Chapter 1

Overview of C
CHAPTER OUTLINE

1.1 The History

1.2 Characters Used in C
1.3 Data Types
1.4 C Tokens
1.5 Structure of a C Program
1.6 printf() and scanf() Functions
1.7 Comments
1.8 Escape Sequence (Backslash Character Constants)
1.9 Operators and Expressions
1.10 Flow of Control
1.11 InputOutput Functions (I/O)
1.12 Arrays
1.13 Structures
1.14 User-defined Data Types
1.15 Unions
1.16 Functions
1.17 Recursion
1.1 THE HISTORY

In 1971 Dennis Ritchi, a system programmer from Bell laboratories,

developed a very powerful language called C for writing UNIX, a large
and complex operating system. Even the compiler of C was written in C.

In fact, it is a high level language that not only supports the necessary
data types and data structures needed by a normal programmer but it
can also access the computer hardware through specially designed
declarations and functions and, therefore, it is often called as a
middlelevel language.
It is popular because of the following characteristics:
1.
2.
3.
4.
5.
6.

Small size
Wide use of functions and function calls
Loose typing
Bitwise low level programming support
Structured language
Wide use of pointers to access data structures and physical memory of the
system

Besides the above characteristics, the C programs are small and efficient.
A C program can be compiled on variety of computers.
1.2 CHARACTERS USED IN C

The set of characters allowed in C consists of alphabets, digits, and

special characters as listed below:
1. Letters: Both upper case and lower case letters of English:
A, B, C, .... X, Y, and Z
a, b, c, .... x, y, and z
2. Decimal digits:
0, 1, 2, .... 7, 8, 9
3. Special characters:
! * + \ <<
#(= ! {
%) ; /
^_ [ :, ?
& _ ] . blank
1.3 DATA TYPES

Every program specifies a set of operations to be done on some data in a

particular sequence. However, the data can be of many types such as a

numbers, characters, floating points, etc. C supports the following simple

data types: Integer data type, character data type, floating point data
type.
1.3.1 Integer Data Type (int)
An integer is an integral whole number without a decimal point. These
numbers are used for counting. Examples of integers are:
923
47
5
15924
56
2245
C allows four types of representation of integers, i.e., integer, long
integer, short integer, and unsigned integer.
1. Integer: An integer is referred to as int. It is stored in one word of the
memory.
2. Long integer: A long integer is referred to as long int or simple long. It is
stored in 32 bits and does not depend upon the word size of the memory.
3. Short integer: A short integer is referred to as short int or simple short. It
is stored in 16 bits and does not depend upon the size of the memory.
4. Unsigned integers: C supports two types of unsigned integers: the
unsigned int and the unsigned short.

The unsigned int is stored in one word of the memory whereas the
unsigned short is stored in 16 bits and does not depend upon the word
size of the memory.
Examples of invalid integers are:
(i) 9, 24, 173

illegal-comma used

(ii) 5.29

illegal-decimal point used

(iii) 79 248

blank used
1.3.2 Character Data Type (char)

It is a non-numeric data type consisting of single alphanumeric character

enclosed between a pair of apostrophes, i.e., single quotation marks.
Examples of valid character type are:

A
N
*
7
It may be noted that the character 7 is different from the numeric value
7. In fact, former is of typechar and later of type int. Each character
has a numeric code, i.e., the ASCII code. For instance, the ASCII code for
the character A is 65 and that of * is 42. A table of ASCII codes is given
in Appendix A.
A character data type is referred to as char. It is stored in one byte of
memory. However, the representation varies from computer to
computer. For instance, some computers support signed as well as
unsigned characters. The signed characters can store integer values from
128 to +127 whereas the unsigned characters store ASCII codes from 0
to 255.
1.3.3 The Floating point (float) Data Type
A floating point number is a real number which can be represented in
two forms: decimal and exponent forms. Floating point numbers are
generally used for measuring quantities.
1. Decimal form: The floating point number in this form has a decimal
point. Even if it is an integral value, it must include the decimal point.
Examples of valid decimal form numbers are:
973.24
849.
73.0
82349.24
9.0004
2. Exponent form: The exponent form of floating point number consists of
the following parts:

<integer>.<fraction>e<exponent>
Examples of valid floating point numbers are:
3.45 e7
0.249 e6

It may be noted that exponent form is a scientific notation wherein the

number is broken into two parts: mantissa and exponent. The mantissa
is a floating point number of decimal form. The exponent part starts with
a letter e followed by an integer (signed or unsigned).
For example, the number 324.5 can be written as 3.245 times 102. In
exponential form, the number is represented as 3.245 e2. In fact, the base
10 has been replaced by the character e (or E).
The utility of exponential form is that very small numbers can be easily
represented by this notation of floating points.
For example, the number 0.00001297 can be written as 0.1297 e 4 or as
12.97 e6 or as 129.7 e7. C allows the following two data types for floating
prints:
1. float: These types of numbers are stored in 32 bits of memory.
2. double: The numbers of double data type are stored in 64 bits of memory.

A summary of C basic data types is given the Table 1.1. From this table, it
may be observed that character and integer type data can also be
declared as unsigned. Such data types are called unsigned data types.
In this representation, the data is always a positive number with range
starting from 0 to a maximum value. Thus, a number twice as big as a
signed number can be represented through unsigned data types.
Table 1.1 Basic data types in C

1.4 C TOKENS

A token is a group of characters that logically belong together. In fact, a

programmer can write a program by using tokens. C supports the
following types of tokens:
1.
2.
3.
4.

Identifiers
Keywords
Constants
Variables

1.4.1 Identifiers
Symbolic names can be used in C for various data items. For example, if
a programmer desires to store a value 27, then he can choose any
symbolic name (say, ROLL) and use it as given below:
ROLL=27;
Where ROLL is a memory location and the symbol = is an assignment
operator.
The significance of the above statement is that ROLL is a symbolic name
for a memory location where the value 27 is being stored. A symbolic
name is generally known as an identifier.
The identifier is a sequence of characters taken from C character set. The
number of characters in an identifier is not fixed though most of the C
compilers allow 31 characters. The rules for the formation of an identifier
are:
1. An identifier can consist of alphabets, digits and and/or underscores.
2. It must not start with a digit.
3. C is case sensitive, i.e., upper case and lower case letters are considered
different from each other.
4. An identifier can start with an underscore character. Some special C names
begin with the underscore.
5. Special characters such as blank space, comma, semicolon, colon, period,
slash, etc. are not allowed.
6. The name of an identifier should be so chosen that its usage and meaning
becomes clear. For example, total, salary, roll no, etc. are self
explanatory identifiers.

Examples of acceptable identifiers are:

TOTAL
Sum
Net_sal
P123
a_b_c
total
_sysreg
Examples of unacceptable identifiers are:
Bas ic (blank not allowed)
H, rent (special character , included)
It may be noted here that TOTAL and total are two different identifier
names.

1.4.2 Keywords
A keyword is a reserved word of C. This cannot be used as an identifier
by the user in his program. The set of C keywords is given in Table 1.2.
Table 1.2 Standard keywords in C

1.4.3 Variables
A variable is the most fundamental aspect of any computer language. It
is a location in the computer memory which can store data and is given a
symbolic name for easy reference. The variables can be used to hold
different values at different times during a program run. To understand
this concept, let us have a look at the following set of statements:
Total = 500.25;

(i)

Net = Totla 100.00;

(ii)

In statement (i), value 500.25 has been stored in a memory location

called Total. The variable Total is being used in statement (ii) for the
calculation of another variable Net. The point worth noting is that the
variable Total is used in statement (ii) by its name not by its value.
Before a variable is used in a program, it has to be defined. This activity
enables the compiler to make available the appropriate amount of space
and location in the memory. The definition of a variable consists of its
type followed by the name of the variable. For example, a variable called
Total of type float can be declared as shown below:
floatTotal;
Similarly, the variable net of type int can also be defined as shown below:
intNet;

Examples of valid variable declarations are:

1.
2.
3.
4.
5.
6.

intcount;
inti,j,k;
charch,first;
floatTotal,Net;
longintsal;
doublesalary.

Let us now look at a variable declaration from a different perspective.

Whenever a variable (say, int val) is declared, a memory location called
val is made available by the compiler as shown in Figure 1.1.

Fig. 1.1 Memory allocated to variable val

Thus, val is the name associated by the compiler to a location in the
memory of the computer. Let us assume that, at the time of execution,
the physical address of this memory location (called val) is 4715 as
shown in Figure 1.2.

Fig. 1.2 The physical address of val

Now, a point worth noting is that this memory location is viewed by the
programmer as a variable called val and by the computer system as an
address 4715. The programmer can store a value in this location with the
help of an assignment operator, i.e., =. For example, if the programmer
desires to store a value 100 into the variable val, he can do so by the
following statement:
val=100;
Once the above statement is executed, the memory location called val
gets the value 100 as shown inFigure 1.3.

C allows the initialization of variables even at the time of declaration as

shown below:
intval=100;
From Figure 1.3, we can see that besides its type a variable has three
entities associated with it, i.e., the name of variable (val), its physical
address (4715), and its contents (100). The content of a variable is also
called its rvalue whereas the physical address of the variable is called
its lvalue. Thus, lvalue and rvalue of variable val are 4715 and 100,
respectively. The lvalue is of more importance because it is an expression
that should appear on the left hand side of assignment operator because
it refers to the variable or object.

Fig. 1.3 The contents of a variable (rvalue)

1.4.4 Constants
A constant is a memory location which can store data in such a manner
that its value during execution of a program does not change. Any
attempt to change the value of a constant will result in an error message.
A constant in C can be of any of the basic data types, i.e., integer
constant, floating point constant, and character constant. const qualifier
is used to declare a constant as shown below:
const <type> <name> = <val>;
where
const: is a reserved word of C
<type>: is any of the basic data types
<name>: is the identifier name
<val>:is the value to be assigned to the constant.
1. Integer constant: It is a constant which can be assigned integer values
only. For example, if we desire to have a constant called rate of type integer
containing a fixed value 50, then the following declaration can be used:
const int rate = 50;

The above declaration means that rate is a constant of type integer having
a fixed value 50. Consider the following declaration:
const int rate;
rate = 50;
The above initialization of constant rate is illegal. This is because of the
reason that a constant cannot be initialized for a value at a place other than
where it is declared. It may be further noted that if a program does not
change or mutates\a constant or constant object then the program is called
as const correct.
2. Floating point constant: It is a constant which can be assigned values of
real or floating point type. For example, if it is desired to have a constant
called Pi containing value 3.1415, then the following declaration can be
used.
constfloatPi=3.1415;
The above declaration means that Pi is a constant of type float having a
fixed value 3.1415. It may be noted here that by default a floating point
constant is of type double.
3. Character constant: A character constant can contain a single character
of information. Thus, data such as Y or N is known as a character
constant. Let us assume that it is desired to have a constant called Akshar
containing the character Q; following declaration can be used to obtain
such a constant.

constcharAkshar=Q;
The above declaration means that Akshar is a constant of type char
having a fixed value Q.
A sequence of characters enclosed within quotes is known as a string
literal. For example, the
character sequence computer is a string literal. When a string literal is
assigned to an identifier declared as a constant, then it is known as
a stringconstant. In fact, a string is an array of characters. Arrays are
discussed later in the chapter.

1.5 STRUCTURE OF A C PROGRAM

A simple program in C can be written as a module. More complex

programs can be broken into submodules. In C, all modules or
subprograms are referred to as functions.
The structure of a typical C program is given below:
main()
{
.;
.;
.;
}
Note:
1. C is a case sensitive language, i.e., it distinguishes between upper case and
lower case characters. Thus, main() is different from Main(). In fact, most
of the characters used in C are lowercase. Hence, it is safest to type
everything in lower case except when a programmer needs to capitalize
some text.
2. Every C program has a function called main followed by parentheses. It is
from here that program execution begins. A function is basically a
subprogram and is complete in itself.
3. The task to be performed by a function is enclosed in curly braces called its
body, i.e., {}. These braces are equivalent to begin and end keywords used in
some other languages like Pascal. The body of function contains a set of
statements and each statement must end with a semicolon.

1.5.1 Our First program

Let us break the myth that C is a difficult language to start with. Without
going into the details, the beginner can quickly start writing the program
such as given below:
#include<stdio.h>
main()
{
puts(Thisismyfirstprogram);
}
The above program displays the following text on the screen:
This is my first program.

1.6 printf() AND scanf() FUNCTIONS

C uses printf() and scanf() functions to write and read from I/O
devices, respectively. These functions have been declared in the header
file called stdio.h.
Let us use printf() function to rewrite our first program. The modified
program is given below:
#include<stdio.h>
main()
{
printf(Thisismyfirstprogram);
}
This program displays the following message on the screen:
This is my first program.
In fact, any text written within the pair of quotes ( ) is displayed as
such by printf() function on the screen. However, this function is
much more powerful than puts() in the sense that it can display all
types of data on the screen as explained below.
1.6.1 How to Display Data using printf() Function
An integer, stored in a variable, can be displayed on the screen by
including a format specifier (%d) within a pair of quotes as shown below:
#include<stdio.h>
main()
{
intage=25;
printf(%d,age);
}
The above program does the following:
1. Declares a variable age of type int.
2. Initializes age to 25, an integer value.
3. Specifies %d, a format specifier indicating that an integer is to be displayed
by printf() and the data is stored in the variable called age.

It may noted that printf() has two arguments separated by a comma,

i.e., format specifier written within quotes and the variable age. It may
be further noted that both messages and format specifiers can be
included within the pair of quotes as shown below:

printf(Theageofstudent=%d,age);
The above statement would display the following data on the screen:
The age of student = 25
Thus, the text between the quotes has been displayed as such but for %d,
the data stored in the variable age has been displayed.
A float can be displayed on the screen by including a format specifier
(%f) within the pair of quotes as shown below:
floatrate=9.5;
printf(Therateofprovidentfund=%f,rate);
The above statements would produce the following display:
The rate of provident fund = 9.5
A char can be displayed by including a format specifier (%c) within the
pair of quotes as shown below:
charch=A
printf(Thefirstalphabetis:%c,ch);
Obviously, the output of the above set of statements would be:
The first alphabet is: A.
The other specifiers which can be used in printf() function are given
in Appendix B.
1.6.2 How to Read Data from Keyboard using scanf()
Similar to printf() function, scanf() function also uses the %d,
%f,%c, and other format specifiers to specify the kind of data to be
read from the keyboard.
However, it expects the programmer to prefix the address operator '&' to
every variable written in the argument list as shown below:
introll;
scanf(%d,&roll);
The above set of statements declares the variable roll of type int, and
reads the value for roll from the keyboard. It may be noted that &, the
address operator, is prefixed to the variable roll. The reason for
specifying & would be discussed later in the book.

Similarly, other format specifiers can be used to input data from the
keyboard.
Example 1: Write an interactive program that reads the marks secured
by a student for four subjectsSub1, Sub2, Sub3, and Sub4, the
maximum marks of a subject being 100. The program shall compute the
percentage of marks obtained by the student.
Solution: The scanf() and prinf() functions would be used to do
the required task. The required program is as follows:
#include<stdio.h>
main()
{
intsub1,sub2,sub3,sub4;
floatpercent;
printf("Enterthemarksforfoursubjects:");
scanf("%d%d%d%d",&sub1,&sub2,&sub3,&sub4);
percent=(sub11sub21sub31sub4)/400.00*100;
printf("Thepercentage5%f",percent);
}
For input data 45 56 76 90, the above program computes the percentage
and displays the following output:
The percentage = 66.750000
Though the output is correct, but it has displayed unnecessary four
trailing zeros. This can be controlled by using field width specifiers
discussed later in the chapter.
1.7 COMMENTS

A comment can be added to the program by enclosing the text between

the pair / * * /, i.e., the pair /* indicates the beginning of the
comment whereas the pair */ marks the end of it. It is also known as
multiple line comment. For example, the following line is a comment:
/*Thisismyfirstprogram*/
Everything written within /*and*/ is ignored by the compiler.
A comment written in this fashion can overflow to multiple lines as
shown below:

/*Thisisanillustrationofmultiplelinecomment.
TheCcompilerignorestheselines.Aprogrammercan
usecommentlinesfordocumentationofhisprograms*/
In fact, a comment is a non-executable statement in the program.
1.8 ESCAPE SEQUENCE (BACKSLASH CHARACTER CONSTANTS)

C has some special character constants called backslash character

constants. These are unprintable ASCII characters which can perform
special functions in the output statements. A backslash character
constant is nothing but a back slash (\) character followed by another
character.
For example, \n can be used in a printf() statement to send the next
output to the beginning of the next line.
Consider the following program:
#include<stdio.h>
main()
{
printf(Abackslashcharacterconstant);
printf(printstheoutputonthenextline.);
}
The output of this program would be:
A backslash character constant prints the output on the next
line.
It may be noted that though two separate printf() statements were
written in the above program, the output has been displayed on the same
line.
In order to display the text in two lines, the \n character constant
should be placed either at the end of the text in the
first printf() statement or at the beginning of the text in the
second printf()statement.
Consider the following modified program:
#include<stdio.h>
main()
{
printf(Abackslashcharacterconstant);
printf(\nprintstheoutputonthenextline.);

}
The character constant \n has been placed at the beginning of the text
of the second printf()statement and, therefore, the output of the
program would be:
A back slash character constant
prints the output on the next line.
Some other important backslash constants are listed below:
Backslash character constant

Meaning

\t

tab

\b

backspace

\a

bell (alert)

\n

newline character

Note:
The backslash characters are also known as escape sequences; such
characters can be used within apostrophes or within double quotes.
1. \t (tab): This character is called a tab character. Whenever it is inserted
in an output statement, the display moves over to the next present tab stop.
Generally, a tab is of 8 blank spaces on the screen. An example of usage of
character \t is given below:
#include<stdio.h>
main()
{
printf(\nhello\tCompBoy);
}
The output of the above statement would be on the new line with spacing
as shown below:
helloCompBoy
2. \b (Backspace): This character is also called backspace character. It is
equivalent to the back space key symbol ()available on the computer or
typewriter. It moves one column backward and positions the cursor on the

character displayed on that column. An example of usage of

character\b is given below:
#include<<stdio.h
main()
{
printf(\nASHOKA\b);
}
The output of the above statement would be:
ASHOK
We can see that the trailing letter A has been overwritten by the
character - at the end of the string constant <ASHOKA> the reason being
that \b moved the curser one column backward, i.e., at A
and printf() printed the character - on that column position.
3. \a (Alert): This character is also called alert or bell character. Whenever
it is inserted in an output statement, it sounds a bell which can be heard by
the user sitting on the computer terminal. It can be used in a situation
where the programmer wants to catch the attention of the user of this
program. An example of usage of \a character is given below:
#include<stdio.h>
main()
{
printf(\nErrorindata\a);
}
The output of the above statement would be the following message on the
screen and a sounding of a bell on the system speaker.
Errorindata
4. \n (new line): As discussed earlier, this character is called newline
character. Wherever it appears in the output statement, the immediate next
output is taken to the beginning of the next new line on the screen. Consider
the statement given below:

#include<stdio.h>
main()
{
printf(\nThisis\natest.);

}
The output of this statement would be:
Thisis
atest.
1.9 OPERATORS AND EXPRESSIONS

An operator is a symbol or letter used to specify a certain operation on

variables in a program. For example, the symbol + is an add operator
that adds two data items called operands.
Expressions: An expression is a combination of operands (i.e.,
constants, variables, numbers) connected by operators and parenthesis.
For example, in the expression given below A and B are operands and +
is an operator.
A+B
C supports many types of operators such as arithmetic, relational,
logical, etc. An expression that involves arithmetic operators is known as
an arithmetic expression. The computed result of an arithmetic
expression is always a numerical value. The expression which involves
relational and/or logical operators is called as a boolean expression or
logical expression. The computed result of such an expression is a logical
value, i.e., either 1 (True) or 0 (False).
The rules of formation of an expression are:
o A signed or unsigned constant or variable is an expression.
o An expression connected by an operator to a variable or a constant is an
expression.
o Two expressions connected by an operator is also an expression.
o Two operators should not occur in continuation.

1.9.1 Arithmetic Operators

The valid arithmetic operators supported by C are given in Table 1.3.
1.9.1.1 Unary Arithmetic Operators
A unary operator requires only one operand or data item. The unary
arithmetic operators supported by C are unary minus (), increment (+
+), and decrement (). As compared to binary operators, the unary
operators are right associative in the sense that they evaluate from right
to left.

Table 1.3 Arithmetic operators

Symbol

Stands for

Example

addition

x+y

subtraction

xy

multiplication

x*y

division

x/y

modulus or remainder

x%y

decrement

x
x

++

increment

x++

The unary minus operator is written before a numerical value, variable

or an expression. Examples of usage of unary minus operator are:
1.
2.
3.
4.
5.

57
2.923
++x
(a*b)
8*((a + b))

It may be noted here that the result of application of unary minus on an

operand is the negation of its operand.
The operators ++ and are unique to C. These are called increment
and decrement operators, respectively. The increment operator ++ adds
1 to its operand. Therefore, we can say that the following expressions are
equivalent.
i=i+1?++i;
For example, if the initial value of i is 10 then the expression ++i will
increment the contents of i to 11. Similarly, the decrement operator --
subtracts 1 from its operand. Therefore, we can say that the following
expressions are equivalent:
j=j1?j;

For example, if the initial value of j is = then the expression --j will
decrement the contents of j to 4.
The increment and decrement operators can be used both as a prefix and
as a postfix to a variable
as shown below:
++xorx++
yory
As long as the increment or decrement operator is not used as part of an
expression, the prefix and postfix forms of these operators do not make
any difference. For example, ++x and x++ would produce the same
result. However, if such an operator is part of an expression then the
prefix and postfix forms would produce entirely different results.
In the prefix form, the operand is incremented or decremented before
the operand is used in the program. Whereas in the postfix form, the
operand is used first and then incremented or decremented.
1.9.2 Relational and Logical Operators
A relational operator is used to compare two values and the result of
such an operation is always logical, i.e., either true or false. The valid
relational operators supported by C are given in Table 1.4.
Table 1.4 Relational operators supported by C
Symbol

Stands for

Example

>

greater than

x>y

>=

greater than equal to

x >= y

<

less than

x<y

<=

less than equal to

x <= y

==

equal to

x==y

!=

not equal to

x!=y

Example 2: Take two variables x and y with initial values 10 and 15,
respectively. Demonstrate the usage of relational operators by
using x and y as operands.
Solution: The usage of relational operators is illustrated below with the
help of a table:
x=10,y=15
Expression

Result

x>y

False

x + 5> = y

True

x<y

True

x<=y

True

x= =y

False

x + 5= =y

True

x! = y

True

A logical operator is used to connect two relational expressions or logical

expressions. The result of such an operation is always logical, i.e., either
true or false. The valid logical operators supported by C are given
in Table 1.5.
Table 1.5 Logical operators supported by C
Symbol

Stands for

Example

&&

Logical AND

x&&y

||

Logical OR

x||y

Logical NOT

|x

Rules of logical operators:

1. The output of a logicalAND operation is true if both of its operands are
true. For all other combinations, the result is false.

2. The output of logicalOR operation is false if both of its operands are

false. For all other combinations, the result is true.
3. The logicalNOT is a unary operator. It negates the value of the operand.

For initial value of x=5 and y=7, consider the following expression:
(x<6)&&(y>6)
The operand x<6 is true and the operand y>6 is also true. Thus,
the result of above given logical expression is also true. However, the
result of following expression is false because one of the operands is
false:
(x<6)&&(y>7)
Similarly, consider the following expression:
(x<6)||(y>7)
The operand x<6 is true whereas the operand y>7 is false. Since
these operands are connected by logical OR, the result of this expression
is true (Rule 2). However, the result of the following expression becomes
false because both the operands are false.
!(x<6)||(y>7)
Note: The expression on the right hand side of logical operators & and ||
does not get evaluated in case the left hand side determines the outcome.
Consider the expression given below:
x&&y
If x evaluates to false (zero), then the outcome of the above expression is
bound to be false irrespective of y evaluating to any logical value.
Therefore, there is no need to evaluate the term y in the above
expression.
Similarly, in the following expression, if x evaluates to true (non zero)
then the outcome is bound to be true. Thus, y will not be evaluated.
x||y

1.9.3 Conditional Operator

C provides a conditional operator (? :) which can help the programmers
in performing simple conditional operations. It is represented by the
symbols ? and ::. For example, if one desires to assign the bigger of the
two variables x and y to a third variable z the conditional operator is an
excellent tool.
The general form of this operation is:
E1?E2:E3
where E1, E2, and E3 are expressions.
In the conditional operation, the expression E1 is tested, if E1 is true then
E2 is evaluated otherwise the expression E3 is evaluated as shown
in Figure 1.4.
Consider the following conditional expression:
z=(x>y)?x:y;
The expression x<y is evaluated. If x is greater than y, then z is
assigned x otherwise z gets y. Examples of valid conditional expressions
are:
1.
y=(x> = 10)?0:10;

Fig. 1.4 The conditional operator

2.
Res = (i<j)? sum + i : sum 1+ j;
3.
q = (a = = 0)?0:(x/y);
It may be noted here that the conditional operator (? :) is also known as a
ternary operator because it operates on three values.

1.9.4 Order of Evaluation of Expressions

A number of logical and relational expressions can be linked together
with the help of logical operators as shown below:

(x<y)||(x>20)&&!(z)||((x<y)&&(z>5))
For a complex expression such as given above, it becomes difficult to
make out as to in what order the evaluation of sub-expressions would
take place.
In C, the order of evaluation of an expression is carried out according to
the operator precedence given in Table 1.6.
Table 1.6 Operator precedence

It may be noted here that in C, false is represented as zero and true as

any non-zero value. Thus, expressions that use relational and logical
operators return either 0 (false) or 1 (true).
1.9.5 Some Special Operators
There are many other operators in C. In this section, the two frequently
used operatorssizeof and comma operatorshave been discussed.
1. Sizeof operator: C provides a compile time unary operator called sizeof
which, when applied on an operand, returns the number of bytes the
operand occupies in the main memory. The operand could be a variable, a
constant or a data type. For example, the following expressions:
a = sizeof (sum);
b = sizeof (char);
c = sizeof (123L);
would return the sizes occupied by variables sum, data type char and
constant 123L on your machine. It may be noted that

1. the parentheses used with sizeof are required when the

operand is a data type with variables or constants, the
parentheses are not necessary.
2. sizeof operator has the same precedence as prefix
increment/decrement operators.

2 Comma operator: The comma operator is used to string together a

number of expressions which are performed in a sequence from left to right.

For example, the following statement

a = (x = 5, x + 2)
executes in the following order
1. value = is assigned to variable x.
2. x is incremented by 2.
3. the value of expression x + 2 (i.e., 7) is assigned to the variable a.

The following points may be noted regarding comma operator:

o A list of expressions separated by a comma is always evaluated from left to
right.
o The final value and type of a list of expressions separated by a comma is
always same as the type and value of the rightmost expression in the list.
o The comma operator has the lowest precedence among all C operator.

1.9.6 assignment Operator

Statements are the smallest executable units of a C program and each
statement is terminated with a semicolon. An assignment statement
assigns the value of the expression on the right hand side to a variable on
the left hand side of the assignment operator (=). Its general form is
given below:
<variablename>=<expression>
The expression on the right hand side could be a constant, a variable or
an arithmetic, relational, or logical expression. Some examples of
assignment statements are given below:
a = 10;
a = b;
a = b*c;
1. The assignment operator is a kind of a store statement, i.e., the value of the
expression on the right hand side is stored in the variable appearing on the
left side of the assignment operator. The variable on the left side of the
assignment operator is also called lvalue and is an accessible address in the
memory. Expressions and constants on the right side of the assignment
operator are called rvalues.

2. The assignment statement overwrites the original value contained in the

variable on the left hand side with the new value of the right hand side.
3. Also, the same variable name can appear on both sides of the assignment
operator as shown below:
count = count + 1;
4. Multiple assignments in a single statement can be used, especially when
same value is to be assigned to a number of variables.
a = b = c = 30;
These multiple assignment statements work from right to left and at the
end, all variables have the same value. The above statement assigns the
value (i.e., 30) to all variables c, b, and a. However, the variables must be of
same type.
5. A point worth nothing is that C converts the type of value on the right hand
side to the data type on the left.

1.9.7 Bitwise Shift Operators

C supports special bitwise shift operators: shift left (<<) and shift right
(>>). We know that data at machine level is represented in the binary
form and it can be shifted to left or right by these special operators.
However, the number of bits to be shifted is specified by an integer
written next to the operatoras shown below:
x << 3;
The above statement means that the value contained in x is shifted to left
by 3 bits. Thus, the leftmost 3 bits of x will be lost and the resultant
vacant 3 rightmost bits will be filled by zeros as shown in Figure 1.5.

Fig. 1.5 The contents of x before and after left shift

Similarly, the following expression would shift the contents of y to right

by two bits:
y >> 2
In this case, the rightmost two bits would be lost and the resultant 2
vacant leftmost bits would be filled by zeros as shown in Figure 1.6.

Fig. 1.6 The contents of y before and after right shift

1.10 FLOW OF CONTROL

A statement is the smallest executable unit of a C program. It is

terminated with a semicolon. It is an instruction given to the computer
to perform a particular task like reading input, displaying output or
evaluating an expression, etc.
A single statement is also called a simple statement. Some examples of
simple statement are given below:
1. int a = 100;
2. S = S + a;
3. count ++;

A statement can also be an empty or null statement as shown below:

;
The high level languages are designed for computers based on VonNeumann architecture. Since this architecture supports only sequential
processing, the normal flow of execution of statements in a high level
language program is also sequential, i.e., each statement is executed in
the order of its appearance in the program. For example in the following
C program segment, the order of execution is sequential from top to
bottom.

The first statement to be executed is x=10, and the second statement

to be executed is y=20. The execution of statement z=x+y will
take place only after the execution of the statement y=20. Thus, the
processing is strictly sequential. Moreover, every statement is executed
once and only once.
Depending upon the requirements of a problem, it is often required to
alter the normal sequence of execution in the program. This means that
we may desire to selectively and/or repetitively execute a program
segment. A number of C control structures are available for controlling
the flow of processing. These structures are discussed in the following
sections.
1.10.1 The Compound Statement
A compound statement is a group of statements separated from each
other by a semicolon. The group of statements, also called a block of
code, is enclosed between a pair of curly braces, i.e., { and }. The
significance of a block is that the sequence of statements enclosed in it is
treated as a single unit. For example, the following group of statements is
a block.
/*ablockofcode*/
{
scanf(%d%d,&a,&b);
c=a+b;
printf(\n%d,c);
}
One compound statement can be embedded in another as shown below:
{
:
{
:
}
:
}

In fact, the function of curly braces { and } in a C program to create a

block, is same as the function ofbegin and end, the reserved words in
Pascal. C calls these braces as delimiters.
1.10.2 Selective Execution(Conditional Statements)
In some cases, it is desired that a selected segment of a program be
executed on the basis of a test, i.e., depending upon the state of a
particular condition being true or false. In C, ifstatement is used
for selective execution of a program segment.
1. The if statement: This statement helps us in the selection of one out of
two alternative courses of action. The general form of ifstatement is
given below:
if(expression)
{
statementsequence
}

where

if:
expression:

statement
sequence:

is a reserved word.
is a boolean expression enclosed within a set of
parentheses (These parentheses are necessary
even if there is a single variable in the
expression).
is either a simple statement or a block. However,
it cannot be a declaration.

Examples of acceptable ifstatements are:

1. if(A>B)A=B;
2. if(total<100){
total=total+val;
count=count+1;
}
3. if((Net>7000)&&(l_tex==500)){
:
}
2 The if-else statement: It can be observed from the above examples that
the simple if statement does nothing when the expression is false. An if-

else statement takes care of this aspect. The general form of this construct
is given below:
if(expression)
{
statementsequence1
}
else
{
statementsequence2
}

where

if:
expression:

is a boolean expression, written within

parentheses.
can be a simple or a compound statement.

statement
sequence1:
else:
statement
sequence2:

is a reserved word.

is a reserved word.
can be a simple or a compound statement.

Examples of ifelse statements are:

1.
2.

if(A>B)C=A;
elseC=B;
if(x==100)
printf(\nEqualto100);
else
printf(\nNotEqualto100);

It may be noted here that both the if and else parts are terminated by
semicolons.
2 Nested if statements (if-else-if ladder): The statement sequence of if
or else may contain another if statement, i.e., the if-else statements can be
nested within one another as shown below:
if(exp1)
if(exp2)
{

:
}
else
if(exp3)
{
:
}
else
{
:
}
It may be noted here that sometimes the nesting may become complex in
the sense that it becomes difficult to decide which if does the else match.
This is called danglingelseproblem. The C compiler follows the
following rule in this regard:
Rule: Each else matches to its nearest unmatched preceding if.
Consider the following nested if:
if(x<50)if(y>5)Net=x+y;elseNet=xy;
In the above statement, the else part matches the second if (i.e., if (y > 5))
because it is the nearest preceding if. It is suggested that the
nested if(s) should be written with proper indentation. The else(s) should
be lined up with their matching if(s). Nested if(s) written in this
fashion are also called if-else-if ladder. For example, the nested if given
above should be written as:
if(x<50)
if(y>5)
Net=x+y;
else
Net=xy;
However, if one desires to match the else with the first if, then the braces
should be used as shown below:
if(x<50){
if(y>5)
Net=x+y;
}
else
Net=xy;

The evaluation of if-else-if ladder is carried out from top to bottom. Each
conditional expression is tested and if found true, only then its
corresponding statement is executed. The remaining ladder is, therefore, by
passed. In a situation where none of the nested conditions is found true, the
final else part is executed.
3 Switch statement (selection of one of many alternatives): If it is
required in a program to select one of several different courses of action,
then the switch statement of C can be used. In fact, it is a multibranch
selection statement that makes the control to jump to one of the several
statements based on the value of an int variable or expression. The general
form of this statement is given as:

switch(expression)
{
caseconstant1:
statement;
break;
caseconstant2:
statement;
break;
:
default:
statement;
}

where
switch:

is a reserved word.

expression:

it must evaluate to an integer or character value.

case:

is a reserved word.

constant:

it must be an int or char compatible value.

statement:

a simple or compound statement.

default:

is a reserved word and is an optional entry.

break:

is a reserved word that stops the execution within the

switch and control comes out of the switch construct.

The switch statement works according to the following rules:

1. The value of the expression is matched with the random case constants of
the switch construct.
2. If a match is found then its corresponding statements are executed and
when break is encountered, the flow of control jumps out of the switch
statement. If break statement is not encountered, then the control continues
across other statement. In fact, switch is the only statement in C which is
error prone. The reason is that if the control is in a particular case and then
it keeps running through all cases in the absence of a proper break
statement. This phenomenon is called fall-through.
3. If no match is found and if a default label is present, then the statement
corresponding to default is executed.
4. The values of the various case constants must be unique.
5. There can be only one default statement in a switch statement.

Examples of acceptable switch statements are:

(i)
switch(BP)
{
case1:total+=100;
break;
case2:total+=150;
break;
case3:total+=250;
}
(ii)
switch(code)
{
case101:Rate=50;break;
case102:Rate=70;break;
case103:Rate=100;break;
default:Rate=95;
}
In the statement (ii), it can be observed that depending on the value of
the code one out of the four instructions is selected and obeyed. For
example, if the code evaluates to 103, the third instruction (i.e., Rate
5100) is selected. Similarly if the code evaluates to 101, then the first
instruction (i.e., Rate = 50) is selected.

1.10.3 Repetitive Execution (Iterative Statements)

Some problems require a set of statements to be executed a number of
times, each time changing the values of one or more variables so that
every new execution is different from the previous one. This kind of
repetitive execution of a set of statements in a program is known as a
loop.
We can categorize loop structures into two categories: nondeterministic loops and deterministicloops. When the number of
times the loop is to be executed is not known, then the loop is called nondeterministic loop otherwise it is called a deterministic loop.
C supports while, do while, and for loop constructs to help repetitive
execution of a compound statement in a program. The while and do
while loops are non-deterministic loops and the for loop is a
deterministic loop.
1. The while loop: It is the fundamental conditional repetitive control
structure in C. The general form of this construct is given below:
while<cond>statement;

where

while:

is a reserved word of C.

<cond>:

is a boolean expression.

statement:

can be a simple or compound statement.

The sequence of operation in a while loop is as follows:

1. Test the condition.

2. If the condition is true, then execute the statement and
repeat step 1.
3. If the condition is false, leave the loop and go on with the
rest of the program.
It may be noted here that the variables used in the <cond> or boolean
expression must be initialized somewhere before the while statement is
encountered. It is a pre-test loop in the sense that the condition is tested
before the body of the loop is executed.
2. The do-while loop: It is another conditional repetitive control structure
provided by C. The syntax of this construct is given below:
do
{

statement;
}
while<cond>;

where

do:

is a reserved word.

statement:

can be a simple or a compound statement.

while:

is a reserved word.

<cond>:

is a boolean expression.

The sequence of operations in a do-while loop is as follows:

1.
2.
3.
4.

Execute the statement.

Test the condition.
If the condition is true, then repeat steps 1 to 2.
If the condition is false, leave the loop and go on with the
rest of the program.

Thus, it performs a post-test in the sense that the condition is not tested
until the body of the loop is executed at least once.
3. The for loop: It is a count controlled loop in the sense that the program
knows in advance as to how many times the loop is to be executed. The
general form of this construct is given below:
for(initialization;expression;increment)
{
statement
}

where

for:

is a reserved word.

initializatio
n:

is usually an assignment expression wherein a loop

control variable is initialized.

expression:

is a conditional expression required to determine

whether the loop should continue or be terminated

increment:

it modifies the value of the loop control variable by a

certain amount.

statement:

can be a simple or a compound statement.

The loop is executed with the loop control variable at initial value, final
value and the values in between.
We can increase the power of for-loop with the help of the comma
operator. This operator allows the inclusion of more than one expression in
place of a single expression in the for statement as shown below:
for(exp1a,exp1b;exp2;exp3a,exp3b;)statement;
Consider the following program segment:
:
for(i=1,j=10;i<=10;i++,j)
{
:
}
The variable i and j have been initialized to values 1 and 10, respectively.
Please note that these initialization expressions are separated by a comma.
However, the required semicolon remains as such. During the execution of
the loop, i increases from 1 to 10 whereas j decreases from 10 to 1
simultaneously. Similarly, the increment and decrement operations have
been separated by a comma in the for statement.
Though the power of the loop can be increased by including more than
one initialization and increment expression separated with the comma
operator but there can be only one test expression which can be simple or
complex.

1.10.3.1 The break and continue Statements

The break statement can be used in any C loop to terminate execution
of the loop. We have already seen that it is used to exit from a switch
statement. In fact, whenever the break statement is encountered in a
loop the control is transferred out of the loop. This is used in a situation
where some error is found in the program inside the loop or it becomes
unnecessary to continue with the rest of the execution of the loop.
Consider the following program segment:
:
while(val!=0)
{
:
printf(\n%d,val);
if(val<0){

printf(\nErrorininput);
break;
}
:
}
Whenever the value contained, in variable val becomes negative, the
message Error in input would be displayed and because of the break
statement the loop will be terminated.
The continue statement can also be used in any C loop to bypass the
rest of the code segment of the current iteration of the loop. The loop,
however, is not terminated. Thus, the execution of the loop resumes with
the next iteration. For example, the following loop computes the sum of
positive numbers in a list of 50 numbers:
:
sum=0;
for(i=0;i<50;i++)
{
printf(\n%d,val);
if(val<=0)continue;
sum=sum+val;
}
:
It may be noted here that continue statement has no relevance as far as
the switch statement is concerned. Therefore, it cannot be used in a
switch statement.
1.10.4 The exit() Function
In the event of encountering a fatal error, the programmer may desire to
terminate the program itself. For such a situation, C supports a function
called exit() which can be invoked by the programmer to exit from the
program. This function can be called from anywhere inside the body of
the program. For normal termination, the programmer can include an
argument 0 while invoking this library function as shown below:
#include<stdio.h>
#include<process.h>
voidmain()
{
:

if(error)exit(0);
:
}
It may be noted here that the file process.h has to be included as header
file because it contains the function prototype of the library
function exit().
1.10.5 Nested Loops
It is possible to nest one loop construct inside the body of another. The
inner and outer loops need not be of the same construct as shown below.
while<cond>
{
do
{
}while<cond>;
for(init;exp;inc)
{
:
}
:
}
The rules for the formation of nested loops are:
1. An outer for loop and an inner for loop cannot have the same control
variable.
2. The inner loop must be completely nested inside the body of the outer loop.

1.10.6 The Goto Statement (Unconditional Branching)

The unconditional transfer of control means that the sequence of
execution will be broken without performing any test and the control will
be transferred to some statement other than the immediate next one. C
supports the following statement for this purpose.
goto<statementlabel>
where goto:

is a reserved word.

<statement

is an identifier used to label the target

label>:

statement.

The goto statement causes control to be transferred to the statement

whose label is specified in thegoto statement. Consider the following
program segment:

gotorpara;

rpara:

The goto statement will cause the control to be transferred to a

statement whose label is rpara. The normal flow of execution will
continue from this statement (i.e., having label rpara) onwards. Consider
the following program segment:
:
k=50;
back:I++
:
sum=sum+k*I;
k;
gotoback;
:
It is clear from the above segment that as and when the control reaches
the goto statement, it is transferred to the statement with label back.
Statement label A statement label is an identifier which can be
placed before any C statement followed by a colon (i.e., :) as shown
below:

again: C = A + B;
Thus, again is a label attached to the statement C = A + B and the
control of execution can be transferred to this statement by the following
goto statement.
gotoagain;
It may be noted here that the transfer of control out of a control
structure is allowed. However, a goto branch into a control structure is
not allowed.
Some example programs using control structures are given below.
Example 3: Write a program that reads three numbers and prints the
largest of them.
Solution: The ifelse ladder would be used. The required program is
given below:
/*Thisprogramdisplaysthelargestofthreenumbers
*/
#include<stdio.h>
main()
{
intA,B,C;
printf(\nEntertheNumbersA,B,C:);
scanf(%d%d%d,&A,&B,&C);
/*ifelseladder*/
if(A>B)
if(A>C)
printf(\nAislargest);
else
printf(\nCislargest);
else
if(B>C)
printf(\nBislargest);
else
printf(\nCislargest);
/*endofladder*/
}

Example 4: Write a program that prints all perfect numbers between 2

and 9999, inclusive. A perfect number is a positive integer such that the
number is equal to the sum of its proper divisors. A proper divisor is any
divisor whose value is less than the number.
Solution: The remainder operator % would be used to find whether a
number is divisor of another or not. The maximum value of perfect
divisor cannot be more than the half the value of the number.
The required program is given below:
/*Thisprogramfindsoutalltheperfectnumbers
between2and9999bothinclusive*/
#include<stdio.h>
main()
{
intnum,divisor;
intsum;
charch;
for(num=2;num,=9999;num++)
{
sum=1;/*1isproperdivisorofallpositive
integers*/
for(divisor=2;divisor<=num/2;divisor++)
{
if(num%divisor550)
sum=sum+divisor;
}
if(sum==num)
printf(\n%disaperfectnumber,num);
}
}
1.11 INPUTOUTPUT FUNCTIONS (I/O)

Most of the programs read and write data through input/output devices
such as screen, keyboard, printer, and disk devices, etc. Obviously, the
programmer has to learn how to perform correct I/O operations so that
efficient programs can be written. We have already used two C
functions printf()and scanf() in the previous examples. In the

following sections, we will discuss some of the following input output

functions:
Function

Type

Description

(1) getchar()

stream

reads a character from a stream

(2) putchar()

stream

writes a character to the stream

(3) getc()

stream

ereads a character from a stream

(4) putc()

stream

writes a character to a stream

(5) gets()

stream

reads a string from a stream

(6) puts()

stream

writes a string to a stream

(7) cgets()

console

reads a string from system consol

(8) getche()

console

reads a character from system consol

(9) putch()

console

writes a character on system consol

(10) cputs()

console

writes a string on system consol

In order to use the above functions, the programmer must include the
stdio.h header file at the beginning of his program. The C or C
environment assumes keyboard as the standard input device and VDU
as standard output device. A consol comprises both keyboard and VDU.
By default, all stream I/Os are buffered. At this point, it is better that we
understand the term buffered I/O before we proceed to discuss the
stream I/O functions.
1.11.1 Buffered I/O
Whenever we read from the keyboard or from a file stored on a device, a
block of data is stored in a buffer called input buffer. The input stream,
therefore, reads from the buffer with the help of a pointer until the input
buffer is empty (see Figure 1.7). As soon as the buffer becomes empty,
the next block of data is transferred into the buffer from the I/O device.

Fig. 1.7 Buffered I/O

Similarly, data written by an output stream pointer is not directly written
onto the device but into an output buffer. As soon as the output buffer
becomes full, a block of data is written on to the I/O device making
output buffer empty.
When the header file stdio.h is included in a program, the following
standard stream pointers are automatically opened:
Name

Meaning

stdin

standard input

stdout

standard output

stderr

standard error

stdaux

auxiliary storage

stdprn

printer

In the computer, stdin and stdout refer to the users console, i.e.,
input device as keyboard and output device as VDU (screen). The stream
pointers stdprn and stdaux refer to the printer and auxiliary
storage, respectively. The error messages generated by the stream are
sent to standard error (stderr).
In order to clear the buffers, a function called fflush() can be used.
For example, input buffer of standard input device can be cleared by
invoking the function: fflush(stdin).
1.11.2 Single Character Functions
C supports many input and output functions based on character. These
functions read or write one character at a time. C supports both stream

and consol I/O character-based functions. We will discuss them in the

following sections.
1. getchar()andputchar()Functions: The getchar() function
reads a single character from the standard input device. For example, if it is
desired to read a character from keyboard in a character variable ch then
the following statements can be used:
:
charch;
ch=getchar();
:
Since the function getchar() is a stream I/O function, it is buffered in
the sense that the character typed by the user is not passed to the variable
ch until the user hits the enter or return key, i.e., . The enter key is itself a
character and it also gets stored in the buffer along with the character typed
by the user. The entry of enter key character into the buffer creates
problems in the normal functioning of getchar() function. Therefore, it is
suggested that after an input operation from the standard input device (i.e.,
keyboard) the input buffer should be cleared. This operation is required to
avoid interference with subsequent input operations. The function call for
this activity is: fflush(stdin). The usage is shown below:
:
charch;
ch=getchar();
fflush(stdin);
:
The function putchar() is used to send a single character to the
standard output device. The character to be displayed on the VDU screen is
included as an argument to the putchar()function as shown below:
:
ch=A;
putchar(ch);
:
Once the above program segment is executed, the character A will be
displayed on the screen.
The putchar() function is also buffered. The function call for clearing
the output buffer is fflush(stdout). The output buffer is cleared on its own
only when a new line character \n is used.

2. getc() and putc() Functions: The getc() and putc()

functions are alsocharacterbasedfunctions. They have been
basically designed to work with files. However, these functions can also be
used to read and write data from standard input and output devices by
specifying stdin and stdout as input and output files, respectively.
For example, the function getc(stdin) is equivalent to the
function getchar(). Both will get a character from the standard input
device (keyboard). The function getchar() by default reads from
keyboard whereas the argument stdin of function getc(stdin) makes it
read from a file represented by the standard input device which is nothing
but the keyboard.
Similarly, the function putc(ch,stdout) is equivalent to the
function putchar(ch). Both the functions send the character ch to the
standard output device, i.e., VDU screen.
3. getche() and putch() Functions: These two functions are characterbased versions of console I/O functions. Therefore, the header file conio.h
has to be included. In fact, these functions act as an extension to the stream
I/O functions. The console I/O functions read from the keyboard and write
to the screen directly.

1. getche(): This function directly reads a character from

the console as soon as it is typed without waiting for the
enter key ( ) to be pressed. There is another similar
functiongetch() which also reads a character from the
keyboard exactly in the same manner but does not show
it on the screen. This function is useful in menu
selections where the programmer does not want to
display the character typed by the user. In fact, the extra
e in the getche() function stands for echo, i.e., display
the character.
Examples of usage of these functions are:
1. ch=getche();
2. ch=getch();
2. putch(): This function directly writes a character
on the console. It takes character as an argument.
Examples of usage of this function are:
1.
2.

putch(\n); //It takes the cursor to next

line of the screen.
putch(ch);//It displays the character stored
in the variable ch on the screen.

3.

putch(A); //It displays the character A on

the screen.

Note: These functions cannot read special keys from the keyboard such
as function keys.
1.11.3 String-based Functions
String-based functions read or write a string of characters at a time. C
supports both stream and console I/O string-based functions.
1. gets() and puts() functions: These functions are string-based I/O
functions. The functiongets() reads a string from the standard input
device (keyboard). It is also buffered and, therefore, the return or enter key
has to be typed to terminate the input. The fflush() function should also be
used after each gets() to avoid interference. For example, the program
segment given below reads a string in a variable name from keyboard.
:
name=gets();
fflush(stdin);
:

Similarly, the function puts() is used to send a string to the standard

output device. The string to be displayed on the VDU screen is included
as an argument to the puts function as shown below:
:
city=NewDelhi;
puts(city);
:
Once this program segment is executed, the following message will be
displayed on the screen:
NewDelhi
1.12 ARRAYS

A detailed discussion on arrays as data structure is given in Chapter 3.

1.13 STRUCTURES

Arrays are very useful for list and table processing. However, the
elements of an array must be of the same data type. In certain situations,
we require a construct that can store data items of mixed data types. C

supports structures for this purpose. The basic concept of a structure

comes from day-to-day life. We observe that certain items are made up
of components or sub-items of different types. For example, a date is
composed of three parts: day, month, and year as shown in Figure 1.8.
Fig. 1.8 Composition of date
Similarly, the information about a student is composed of many
components such as name, age, roll number., and class; each belonging
to different types (see Figure 1.9).

Fig. 1.9 Student

The collective information about the student as shown above is called a
structure. It is similar to a record construct supported by other
programming languages. Similarly, the date is also a structure. The term
structure can be precisely defined as a group of related data items of
arbitrary types. Each member of a structure is, in fact, a variable that
can be referred to through the name of the structure.
1.13.1 Defining a Structure in C
A structure can be defined in C by the keyword struct followed by its
name and a body enclosed in curly braces. The body of the structure
contains the definition of its members and each member must have a
name. The declaration ends by a semicolon. The general format of
structure declaration in C is given below:
struct<name>{
member1
member2
:
membern
};
where

struct:

is the keyword.

<name>:

is the name of the structure.

member1,2n:

are the individual member declarations.

Let us now define the following C structure to describe the information

about the student given inFigure 1.9.
structstudent
{
charName[20];
intage;
introll;
charclass[5];
};
The above declaration means that the student is a structure consisting of
data members: Name, age, roll, and class. A variable of this type can be
declared as given below:
structstudentx;
Once the above declaration is obeyed, we get a variable x of type student
in the computer memory which can be visualized as shown in Figure
1.10.
The programmer is also allowed to declare one or more structure
variables along with the structure declaration as shown below:

Fig. 1.10 Memory space allocated to variable x of type student

structstudent
{
charname[20];
intage;
introll;
charclass[5];
}
stud1,stud2;

In the above declaration, the structure declaration and the variable

declaration of the structure type have been clubbed, i.e., the structure
declaration is followed by two variable names stud1andstud2 and
the semicolon.
1.13.2 Referencing Structure Elements
The structure variable x has four members: name, age, roll, and class.
These members can be designated as: x.name, x.age, x.roll, and x.class,
respectively. In this notation, an individual member of the structure is
designated or qualified by the structure variable name followed by a
period and the member name. The period is called a structure member
operator or simply a dot operator.
However, the information of a student (Figure 1.10) can be stored in
structure variable x in either of the following ways:
1. Initialize the elements of the structure
strcpy(x.name,SACHINKUMAR);
x.age=18;
x.roll=10196;
strcpy(x.class,CE41);
2. The variable x can be read in a C program by the I/O statements as shown
below:
gets(x.name);
ffush(stdin);
scanf(%d%d,&xage.,&x.roll);
gets(x.class);

1.13.3 Arrays of Structures

An array of structures can be declared just like an ordinary array.
However, the structure has to be defined before an array of its type is
declared. For example, a 50-element array of type student called
stud_list can be defined as shown below:
structstudent{
charname[20];
introll;
intsub1,sub2,sub3,sub4;
inttotal;
};
structstudentstud_list[50];/*declarationofan
arrayofstructures*/

1.13.4 Initializing Structures

When a structure variable is declared, its data members are not
initialized and, therefore, contain undefined values. Similar to arrays,
the structure variables can also be initialized. For instance, the structure
shown in Figure 1.10 can be declared and initialized as shown below:
structstudent
{
charname[20];
intage;
introll;
charclass[5];
};
/*Initializevariableoftypestudent*/
structstudentstud1={SACHINKUMAR,18,10196,
CE41};
Please notice that the values are enclosed within curly braces. The values
are assigned to the members of the structure in the order of their
appearance, i.e., first value is assigned to the first member, second to
second, and so on.
1.13.5 Assignment of Complete Structures
It has been appreciated that the individual fields of a structure can be
treated as simple variables. The main benefit of a structure, however, is
that it can be treated as a single entity. For example, if two structure
variables stud1 and stud2 have been declared as shown below
studentstud1,stud2;
then the following assignment statement is perfectly valid.
stud1=stud2;
This statement will perform the necessary internal assignments. The
concept of structure assignment can be better understood by the
following example program. In this program, a structure variable stud1 is
initialized with certain values and it is assigned to another structure
variable stud2. The contents of the variable stud2 are displayed.

/*Thisprogramillustratescopyingofcomplete
structures*/
#include<stdio.h>
main()
{
structstudent{
charname[20];
introll;
intage;
charclass[6];
};
/*initializeavariableoftypestudent*/
structstudentstud1={SachinKumar,101,16,
CE_IV};
structstudentstud2;
/*assignstud1tostud2*/
stud2=stud1;
/*Displaycontentsofstud2*/
printf(\nThecopiedcontentsare....);
printf(\nName:%s,stud2.name);
printf(\nRoll:%d,stud2.roll);
printf(\nAge:%d,stud2.age);
printf(\nclass:%s,stud2.class);
}
1.13.6 Nested Structures
Nested structures are structures as member of another structure. For
instance, the date of birth (DOB) is a structure within the structure of a
student as shown in figure below. These types of structures are known as
nested structures.

The nested structure shown in Figure 16.4 can be declared in a program

as shown below:
/*Nestedstructures/*
structdate
{
intdd;
intmm;

intyy;
};
structstudent
{
charname[20];
introll;
structdatedob;
intmarks;
};
It may be noted here that the member of a nested structure is referenced
from the outermost to inner most with the help of dot operators. Let us
declare a variable stud of type student:
structstudentstud;
The following statement illustrates the assignment of value 10 to the
member mm of this nested structure stud:
stud.dob.mm=10;
The above statement means: store value 10 in mm member of a structure
dob which itself is a member of structure variable called stud.
1.14 USER-DEFINED DATA TYPES

In addition to the simple data types such as integer,float,char,

etc., C allows the users to declare new data types called user-defined
data types. These data types can be declared with the help of a keyword
called typedef whose general form is given below:
typedef <type> <new_type>
where

typedef:

is a reserved word.

<type>:

is an existing data type.

<new_type>:

is the desired user-defined data type.

For example, consider the following declaration:

typedeffloatxyz;

For example, consider the following declaration:

typedeffloatxyz;
In this declaration, an existing data type (float) has been redefined
as xyz. Now, the user can usexyz as a synonym for the typefloat as
a new type as shown below.
xyzx1,x2;
In fact, the above declaration is equivalent to the following declaration:
floatx1,x2;
Apparently, it looks as if we have not achieved anything great because
the user will have to rememberxyz as a new type whereas its equivalent
float type is already available. Nevertheless, the utility of typedef lies in a
situation where the name of a particular type is very long. The user can
also self document his code as shown in the following declaration:
typedeffloatsalary;:
salarywagesofmonth;
In the above declaration, the variable wages-of-month has been declared
of type salary, where salary itself is of float type. Thus, we can see that
readability of a program can be enhanced by self documenting the
variables and their data types with the help of typedef.
For example, the declaration:
typedefintage;
can be used to increase the clarity of code as shown below :
ageboy,girl,child;
The above statement clearly suggests as to what data the variables boy,
girl, and child are going to store. Similarly, typedef can also be used to
define arrays.
typedeffloatlists[20];
listsBP,SAL,GSAL;

These statements are equivalent to the following:

floatBP[20],SAL[20],GSAL[20]
typedef is also extremely useful for shortening the long and
inconvenient declarations of structures. For instance, the structure struct
can be declared as a typedef as shown below:
typedefstruct{
charname[20];
intage;
introll;
charclass[6];
}student;
Now the structure student has become the type and a variable of this
type can be defined as:
studentstud1,stud2;
The above declaration means that stud1andstud2 are variables of
type student.
1.14.1 Enumerated Data Types
A user can define a set of integer type identifiers in the form of new data
types called enumerated data types. For instance, the user can create a
new data type called color with the help of enum definition as shown
below:
enumcolor{
Blue,
Yellow,
White,
Black
};
where enum is the keyword that defines an enumerated data type.
The enum declaration given above means that a new type, color, has been
defined. This type contains the constant values: Blue, Yellow, White, and
Black. Each of these symbols stands for an integer value. By default, the
value of the first symbol (i.e., Blue) is 0, the value of the second symbol is

1, and so on. Once the type color has been defined, then variables of that
type can be defined in the following manner:
enumcolorA,B;
Now, the variable A of type color can use the enumerations (blue, yellow,
etc.) in a program.
Thus, the following operations on variable A are valid:
A=red;
:
if(A==red)
x=x+10;
else
x=x5;
Enumerated data type can be precisely defined as the ordered set of
distinct constant values defined as a data type in a program.
Examples of some valid enumerated data types are:
1. enumfurniture
{
chair,
table,
stool,
desk
};
2. enumcar
{
Maruti,
Santro,
Micra,
Indigo
}

It may be noted here that the ordering of elements in enumerated data

type is specified by their position in the enum declaration. Thus for
furniture type, the following relations hold good:
Expression

Chair < Table

True

Stool < Desk

True

Table == Desk

False

Chair < Desk

True

The enumerated types can also be used in a switch construct. Consider

the following program segment.
:
/*Enumeratedtypedeclaration*/
enumfurniture
{
chair,
table,
stool,
desk
};
main()
{
/*Enumeratedvariabledeclaration*/
enumfurnitureitem;
floatcost;
:
switchitem
{
casechair:cost=250;break;
casetable:cost=700;break;
casestool:cost=120;break;
casedesk:cost=1500;break;
}
:
Similarly, enumerated data types can also be used as a control variable in
a for loop as shown below:
:
for(item=chair;item<=desk;item++)
{
:
:
}

The enumerations are also called symbolic constants because an

enumeration has a name and an integer value assigned to it which
cannot be changed in the program. Therefore, enumerated data types
cannot be plainly used in I/O statements. As a result, the
statements scanf,printf,gets,puts, etc. cannot be directly
applied to these data types because the results will not be as desired.
1.15 UNIONS

Suppose in a fete, every visitor is assured to win any one of the items
listed below:
1.
2.
3.
4.

Wall clock
Toaster
Electric Iron
DVD player

Obviously, a visitor to the fete must carry a bag large enough to carry the
largest item so that any one of the items won by him can be held in that
bag. With the same philosophy in mind, the designers of C introduced a
special variable called union which may hold objects of different sizes
and types but one at a time. Though a union is a variable and can hold
only one item at a time, it looks like a structure. The general format of a
union is given below:
union<name>{
member1
member2
:
member
};
where

union:

is a keyword.

<name>:

is the name of the union.

member1,2,n:

are the individual member declarations.

1.16 FUNCTIONS

A function is a subprogram that can be defined by the user in his

program. It is a complete program in itself in the sense that its structure

is similar to main() function except that the name main is replaced by

the name of the function. The general form of a function is given below:
<type> <name> (arguments)
wher
e

<type>:

is the type of value to be returned by the

function. If no value is returned, then keyword
void should be used.

<name>:

is a user-defined name of the function. The

function can be called from another function
by this name.

arguments
:

is a list of parameters (parameter is the data

that the function may receive when called from
another function). This list can be omitted by
leaving the parameters empty.

The program segment enclosed within the opening brace and closing
brace is known as the function body. For example, the main function in C
is written as shown below:
main()
{
}
Let us write a function tri_area()which calculates the area of a
triangle. Two variables base and height will be used to read the values
from the keyboard. The function is given below:
floattri_area()
{
floatbase,height,area;
printf(\nEnterBaseandHeight);
scanf(%f%f,base,height);
area=(0.5)*base*height;
return(area);
}
It may be noted here that the function tri_area() has defined its own
variables. These are known as local variables. The advantage of their

usage is that once the execution of the function is over these variables are
disposed off by the system, freeing the memory for other things.
Let us write another function clear() which clears the screen of the
VDU
/*Thisfunctionclearsthescreenbywritingblank
linesonthescreen*/
voidclear()
{
inti;
for(i=0;i<=30;i++)
printf(\n);
}
Similarly, let us write a function add which receives two parameters x
and y of type integer from the calling function and returns the sum of x
andy.
intadd(intx,inty)
{
inttemp;
temp=x+y;
returntemp;
}
In the above function, we have used a local variable temp to obtain the
sum of x and y. The value contained in temp is returned through return
statement. In fact, variable temp has been used to enhance the
readability of the function only, otherwise it is unnecessary. The function
can be optimized as shown below:
indadd(intx,inty)
{
return(x+y);
}
Thus, a function has following four elements:
1.
2.
3.
4.

A name
A body
A return type
An argument list

The name of the function has to be unique so that the compiler can
identify it. The body of the function contains the C statements which
define the task performed by the function. The return type is the type of
value that the function is able to return to the calling function. The
argument or parameter list contains the list of values of variables from
outside required by the function to perform the given task. For example,
the function given above has a name add. It has a body enclosed between
the curly braces. It can return a value of type int. The argument list is x
and y.
1.16.1 Function prototypes
Similar to variables, all functions must be declared before they are used
in a program. C allows the declaration of a function in the calling
program with the help of a function prototype. The general form of a
function prototype in the calling program is given below:
<type> <name> (arguments);
wher
e

<type>:

is the type of value to be returned by the

function.

<name>:

is a user-defined name of the function.

arguments
:

is a list of parameters.

;:

the semicolon is necessary.

For example, the prototype declarations for the functions tri_area(),

clear(),andadd() can be written as shown below:
voidtri_area();
voidclear();
intadd(intx,inty);
It may be noted that a function prototype ends with a semicolon. In fact,
the function prototype is required for those functions which are intended
to be called before they are defined. In absence of a prototype, the
compiler will stop at the function call because it cannot do the type
checking of the arguments being sent to the function. Therefore, the
main purpose of function prototype is to help the compiler in static type
checking of the data requirement of the function.

1.16.2 Calling a Function

A function can be called or invoked from another function by using its
name. The function name must be followed by a set of actual parameters,
enclosed in parentheses and separated by commas. A function call to
function tri_area from the main program can be written as:
tri_area();
This statement will transfer the control to function tri_area. Since
this function does not require any parameter, the argument list is empty.
Similarly, in order to call the function clear, the following statement will
be sufficient:
clear();
A complete C program involving the two functions tri_area and clear
is given below:
#include<stdio.h>
/*Functionptototypes*/
floattri_area();
voidclear();
main()
{
charch;
intflag;
floatArea;
clear();/*Clearthescreen*/
flag=0;
while(flag==0)
{
Area=tri_area();/*calltri_areatocomputearea
*/
printf(\nAreaofTriangle=%5.2f,Area);
printf(\nWanttocalculateagain:EnterY/N);
fflush(stdin);
ch=getchar();
if(ch==N||ch==n)
flag=1;
clear();

}
}
floattri_area()
{
floatbase,height,area;
printf(\nEnterBaseandHeight);
scanf(%f%f,&base,&height);
area=(0.5)*base*height;
return(area);
}
/*Thisfunctionclearsthescreenbywritingblank
linesonthescreen*/
voidclear()
{
inti;
for(i=0;i<=30;i++)
printf(\n);
}
It may be noted here that when this program is executed, the first
statement to be executed is the first executable statement of the
function main() of the program (i.e., clear() in this case). A function
will be executed only after it is called from some other function or itself.
The variables base, height, and area are local to function tri_area and
they are not available in themain() function. Similarly, the variable flag
declared in function main() is also not available to
functions tri_area() and clear().
However, the variables declared outside any function can be accessed by
all the functions. Such variables are called global variables.
Let us consider the following program which declares the variables flag
and area outside the main function instead of declaring them inside the
functions main() and tri_area(), respectively.
#include<stdio.h>
/*Functionprototypes*/
voidtri_area();
voidclear();
/*GlobalVariables*/
intflag;
floatarea;

main()
{
charch;/*LocalVariable*/
clear();/*Clearthescreen*/
flag=0;
while(flag==0)
{
tri_area();/*calltri_areatocomputearea*/
printf(\nAreaofTriangle=%5.2f,area);
printf(\nWanttocalculateagain:EnterY/N);
fflush(stdin);
ch=getchar();
if(ch==N||ch==n)
flag=1;
clear();
}
}
/*FunctionDefinitions*/
voidtri_area()
{
floatbase,height;
printf(\nEnterBaseandHeight);
scanf(%f%f,&base,&height);
area=(0.5)*base*height;
}
/*Thisfunctionclearsthescreenbywritingblank
linesonthescreen*/
voidclear()
{
inti;
for(i=0;i<=30;i++)
printf(\n);
}
It may be noted in the above program that the variable area is available
to both the functions tri_area() and main(). On the other hand,
the local variables base and height are available to
functiontri_area only. They are neither available to main() function
nor to the function clear(). Similarly, variable i is also local to
function clear().

In our previous programs, we have used our own function clear() to

clear the screen. C also provides a function called clrscr() that clears
the screen. From now onwards, we will use this function. However, in
order to use clrscr() in a program we will have to include conio.h, a
header file at the beginning of the program.
Note: We should not use global variables in our programs as they make
the program or function insecure and error prone.
1.16.3 Parameter Passing in Functions
We have seen that two-way communication between functions can be
achieved through global variables. This technique has its own limitations
and drawbacks. Let us now see how this communication between the
calling function and the called function can be improved and made
fruitful.
The two-way communication between the various functions can be
achieved through parameters and return statement as done in previous
section. The set of parameters defined in a function are
calledformal or dummy parameters whereas the set of
corresponding parameters sent by the calling function are called actual
parameters. For instance, the variables xandy in the argument list
of function add() are formal parameters. The parameters included in
the function call, i.e., val1andval2 of main() are actual
parameters.
The function that receives parameters can be called in one of the
following two ways:
1. Call by value.
2. Call by reference.
1. Call by value: For better understanding of the concept of parameter
passing, let us consider the function big() given below. It receives two
values as dummy arguments xandy and returns the bigger of the two
through a return statement.
/*Thisfunctioncomparestwovariablesandreturnsthe
biggerofthetwo*/
intbig(intx,inty)
{
if(x>y)
returnx;
else
returny;
}

The functionbig() can be called from the functionmain() by

including the variables a and b as the actual arguments of the function call
as shown below.
main()
{
inta=30;
intb=45;
intc;
c=big(a,b);
printf(\nThebigger=%d,c);
}
The output of the above program would be:
Thebigger=45
Let us now write a program which reads the values of two variables a and
b. The values of a and b are sent as parameters to a function
exchange(). The function exchanges the contents of the parameters with
the help of a variable temp. The changed values of a and b are printed.
#include<stdio.h>
#include<conio.h>
voidexchange(intx,inty);
main()
{
inta,b;
clrscr();/*clearscreen*/
printf(\nEntertwovalues);
scanf(%d%d,&a,&b);
exchange(a,b);
/*printtheexchangedcontents*/
printf(\nTheexchangedcontentsare:%dand%d,
a,b);
}
voidexchange(intx,inty)
{
inttemp;
temp=x;
x=y;
y=temp;
}

The above program seems to be wonderful but it fails down badly because
variables a and b do not get any values back from the function
exchange(). The reason is that although all values of actual parameters
(i.e., a and b) have been passed on to formal parameters (i.e., x and y)
correctly, the changes done to the values in the formal parameters are not
reflected back into the calling function.
Once the above program is executed for input value (say 30 and 45), the
following output is produced:
The exchanged contents are: 30 and 45
The reason for this behaviour is that at the time of function call, the
values of actual parameters a and b get copied into the memory locations of
the formal parameters x and y of the functionexchange().
It may be noted that the variables x and y have entirely different memory
locations from variablesaandb. Therefore any, changes done to the
contents of x and y are not reflected back to the variables a and b. Thus, the
original data remains unaltered. In fact, this type of behaviour is desirable
from a pure function.
Since in this type of parameter passing, only the values are passed from
the calling function to the called function, the technique is called call by
value.
2. Call by reference: If the programmer desires that the changes made to
the formal parameters be reflected back to their corresponding actual
parameters, then he should use call by reference method of parameter
passing. This technique passes the addresses or references of the actual
parameters to the called function. Let us rewrite the
function exchange(). In fact, in the called function, the formal arguments
receive the addresses of the actual arguments by using pointers to them.

For instance, in order to receive the addresses of actual parameters of int

type the exchange()function can be rewritten as shown below:
voidexchange(int*val1,int*val2)
{
inttemp;
temp=*val1;
*val1=*val2;
*val2=temp;
}
It may be noted that the above function has used two pointers
called val1andval2 to receive the addresses of two int type actual
parameters. However, initially, these are dangling pointers as shown
below:

The exchange() function would be called from the main() function by

sending the addresses of the actual arguments as shown below:
exchange (&a,&b);
After the exchange() function is called by the above statement, the
addresses of the actual parameters a and b would be assigned to
pointers val1 and val2, respectively as shown below:

Since the function exchanges the contents of variables being pointed by

the pointers val1 and val2, the contents of variables a and b get
exchanged which is the goal of this exercise.
The complete program is given below:
#include<stdio.h>
#include<conio.h>
voidexchange(int*val1,int*val2);
main()
{
inta,b;
clrscr();/*clearscreen*/
printf(\nEntertwovalues);
scanf(%d%d,&a,&b);
exchange(&a,&b);
/*printtheexchangedcontents*/
printf(\nTheexchangedcontentsare:%dand%d,
a,b);
}
voidexchange(int*val1,int*val2)
{
inttemp;
temp=*val1;
*val1=*val2;
*val2=temp;
}

After the above program is executed for input values (say 30 and 45), the
following output is produced:
Theexchangedvaluesare:4530
Now, we can see that the program has worked correctly in the sense that
changes made inside the function exchange() have been available to
corresponding actual parameters a and b in the function main(). Thus,
the address operator & has done the required task.
The main advantage of call by reference method is that more than one
value can be received back from the called function.
Note:
1. As far as possible, we should use pass by value technique of parameter
passing. If some value is to be received back from the function, it should be
done through a return statement of the function.
2. A function call can use both the methods, i.e., call by value and call by
reference simultaneously.For example, the following is a valid function
prototype:
voidcalc(inta,float*b,float*c);
3. When actual parameters are passed from the calling function to the called
function by call by value method, the names of the corresponding dummy
parameters could be the same or different. However, it is suggested that
they may be named different in order to make the program more elegant.
4. Array parameters: Arrays can also be passed as parameters to functions.
This is always done through call by reference method. In C, the name of the
array represents the address of the first element of the array and, therefore,
during the function call only the name of the array is passed from the calling
function. For example, if an array xyz of size 100 is to be passed to a
functionsome() then the function call will be as given below:,

some(xyz);
The formal parameters can be declared in the called function in three
ways. The different ways are illustrated with the help of a
function some() that receives an array xyz of size 100 of integer type.
Method I:
In this method, the array is declared without subscript, i.e., of an known
size as shown below:

voidsome(intxyz[])
{
}
The C compiler automatically computes the size of the array.
Method II:
In this method, an array of same size and type is declared as shown
below:
voidsome(intxyz[100])
{
}
Method III:
In this method, the operator * is applied on the name of the array
indicating that it is a pointer.
voidsome(int*xyz)
{
}
All the three methods are equivalent and the function some() can be
called by the following statement:
some(xyz);
Note:
1. The arrays are never passed by value and cannot be used as the return type
of a function.
2. In C, the name of the array represents the address of the first element or the
zeroth element of the array. So when an array is used as an argument to a
function, only the address of the array gets passed and not the copy of the
entire array. Hence, any changes made to the array inside the function are
automatically reflected in the main function. You do not have to
use ampersand(&) with array name even though it is a call by reference
method. The ampersand(&) is implied in the case of array.

1.16.4 Returning Values from Functions

Earlier in this chapter, we have used functions which either return no
value or integer values. The functions that return no value are declared
as void. We have been prefixing int to functions which return integer
values. This is valid but unnecessary because the default return value of a
function in C is of type int. Thus, the following two declarations are
equivalent in the sense that we need not prefix int to a function which
returns an integer value.
intadd(inta,intb);
add(inta,intb);
On the other hand, whenever a function returns non-integer values, the
function must be prefixed with the appropriate type. For example, a
function xyz that returns a value of type float can be declared as shown
below:
floatxyz(.....)
{floata;
:
returna
}
Similarly, a function ABC that returns a value of type char can be
declared as shown below:
charABC(.....)
{charch;
:
returnch;
}
1.16.5 passing Structures to Functions
Similar to variables, the structures can also be passed to other functions
as arguments both by value and by reference.
1. Call by value: In this method of passing the structures to functions, a copy
of the actual argument is passed to the function. Therefore, the changes
done to the contents of the structures inside the called function are not
reflected back to the calling function.
The following program uses a structure variable called stud of type student
and passes the structure variable by value to a

function print_data(). The function print_data() displays the

contents of the various members of the structure variables.
/*Thisprogramillustratesthepassingofthe
structuresbyvaluetoafunction*/
#include<<stdio.h>>
structstudent
{
charname[20];
intage;
introll;
charclass[5];
};
/*Prototypeofthefunction*/
voidprint_data(structstudentSob);
main()
{
structstudentstud;
printf(\nEnterthestudentdata);
printf(\nName:);fflush(stdin);
gets(stud.name);
printf(\nAge:);scanf(%d,&stud.age);
printf(\nRoll:);scanf(%d,&stud.roll);
printf(\nClass:);fflush(stdin);
gets(stud.class);
print_data(stud);/*Thestructureisbeingpassedby
value*/
}
voidprint_data(structstudentSob)
{
printf(\nThestudentdata..);
printf(\nName:);fflush(stdout);
puts(Sob.name);
printf(Age:%d,Sob.age);
printf(\nRoll:%d,Sob.roll);
printf(\nClass:);fflush(stdout);
puts(Sob.class);
}
2. Call by reference: In this method of passing the structures to functions,
the address of the actual structure variable is passed. The address of the
structure variable is obtained with the help of address operator . Thus, the
changes done to the contents of such a variable inside the function are

reflected back to the calling function. The program given below illustrates
the usage of call by reference method.
This program reads the data of a student in the function main() and passes
the structure by reference to the function called print_data().
/*Thisprogramillustratesthepassingofthe
structuresbyreferencetoafunction*/
#include<stdio.h>
structstudent
{
charname[20];
intage;
introll;
charclass[5];
};
/*Prototypeofthefunction*/
voidprint_data(structstudent&Sob);
main()
{
structstudentstud;
printf(\nEnterthestudentdata);
printf(\nName:);fflush(stdin);
gets(stud.name);
printf(\nAge:);scanf(%d,&stud.age);
printf(\nRoll:);scanf(%d,&stud.roll);
printf(\nClass:);fflush(stdin);
gets(stud.class);
print_data(&stud);/*Thestructureisbeingpassedby
reference*/
}
voidprint_data(structstudent&Sob)
{
structstudent*ptr;
ptr=sob;
printf(\nThestudentdata..);
printf(\nName:);fflush(stdout);
puts(ptr>name);
printf(Age:%d,ptr>age);
printf(\nRoll:%d,ptr>roll);
printf(\nClass:);fflush(stdout);
puts(ptr>class);
}

It may be noted here that structures are usually passed by reference in

order to prevent the overheads associated with the technique of passing
structures by value. The overheads are extra memory space and CPU time
used to pass bigger structures.
3. Returning structures from functions: It is possible to supply a
structure as a return value of a function. This can be done by specifying
structure as the return type of the function. This feature of C is helpful in a
situation where a structure has been passed by value to a function and the
changes done to the contents of the structure are needed by the calling
function without disturbing the original contents. This concept is illustrated
in the example given below:

Example 5: Write a program that reads the data of a students whose

structure is given in the figure shown below. The structure is then passed
by value to a function called changedata() where the name of the
student is capitalized. The changed contents of the student data are
returned by the function to the function main().

Solution: We will use the function toupper() in this program to

capitalize the name of the student inside the function change
data(). The required program is given below:
/*Thisprogramillustratestheprocessofreturning
astructurefromafunction*/
#include<stdio.h>
#include<ctype.h>
structstudent
{
charname[20];
intage;
introll;
charclass[5];
};
/*Prototypeofthefunction*/
voidprint_data(structstudent*ptr);
structstudentchange_data(structstudentSob);
main()
{

structstudentstud,newstud;
printf(\nEnterthestudentdata);
printf(\nName:);fflush(stdin);
gets(stud.name);
printf(\nAge:);scanf(%d,&stud.age);
printf(\nRoll:);scanf(%d,&stud.roll);
printf(\nClass:);fflush(stdin);
gets(stud.class);
/*callfunctiontochangedata*/
newstud=change_data(stud);
printf(\nThechangeddata..);
print_data(&newstud);/*Thestructureisbeing
passedbyvalue*/
}
/*Thisfunctionchangesthedataofa
student*/
structstudentchange_data(structstudentSob)
{
inti=0;
while(Sob.name[i]!=\0)
{/*Capitalizetheletters*/
Sob.name[i]=toupper(Sob.name[i]);
i++;
}
return(Sob);
}
voidprint_data(structstudent*ptr)
{
printf(\nThestudentdata..);
printf(\nName:);fflush(stdout);
puts(ptr>name);
printf(Age:%d,ptr>age);
printf(\nRoll:%d,ptr>roll);
printf(\nClass:);fflush(stdout);
puts(ptr>class);
}
1.17 RECURSION

We have already discussed iteration through various loop structures

discussed in Section 1.10.3. Iteration can be defined as an act of

performing computation each time by the same method and the result of
computation is utilized as the source of data in the next repetition of the
loop. For example the factorial of an integer N can be computed by the
following iterative loop:
fact=1;
for(i=1;i<=N;i++)
fact=fact*i;
printf(\nThefactorialof%d%d,N,fact);
We can also use an alternative approach wherein we reduce the problem
into a smaller instance of the same problem. For example, factorial of N
can be reduced to a product of N and the factorial of N 1 as given
below:
Fact ( N )= N Fact (N 1)
In the above statement, a function Fact(N) has been defined in terms of
Fact(N 1), where Fact (N 1) is a smaller problem as compared to
Fact(N). This type of definition is known as recursive
definition.Recursion can be defined as: the ability of a concept being
defined within the definition itself. In programming terms, recursion is
defined as the ability of a function being called from within the function
itself.
Now, by the same definition, the factorial of (N 1) can be defined as
given below :
Fact(N 1) = (N 1) Fact(N 2)
We can continue this process till we end up with Fact(0) which is equal
to 1, and is also the terminating condition for this reduction process. The
terminating condition for a recursive function is also called the basic
solution. For instance, the basic solution for a list of elements is that if
the list contains only one element then this element is largest as well as
smallest element. Similarly, if a list contains only one element then it is
already sorted.
The process described above can be implemented by using a recursive
function Fact() which is defined in terms of itself as shown below:

In C such recursive functions can be implemented by making the

function call itself. For example, the recursive function Fact(N) can be
defined as follows:

intfact(intN)
{
if(N==0)
return(1);
else
return(N*fact(N1);
}
It may be further noted that the function Fact is being called by itself but
with parameter N being replaced by N 1. A trace of Fact (N) for N = 4 is
given in Fig.1.11

Fig. 1.11 Trace of Fact(4)

Example 7: Write a program that uses recursive function fact (N) to
compute the factorial of a given number N.
Solution: A complete program to compute factorial of a number using
recursion is given below.
/*Aprogramtocomputefactorialofanumberusing
recursion*/
#include<stdio.h>
#include<conio.h>
longintfact(int);
voidmain()
{
intnum,factorial;
clrscr();

printf("\nThisprogramcomputesthefactorialofa
number");
printf("\nEnterthenumber:");
scanf("%d",&num);
factorial=fact(num);
printf("\nThefactorialof%dis%d",num,
factorial);
}
longintfact(intn)
{
if(n==0)
return(1);
else
return(n*fact(n1));
}
Recursion can be used to write simple, short, and elegant programs.
However, a recursive algorithm requires a basic condition that must
terminate the recursive calls. The absence of this condition would result
in infinite recursive calls. For example, in function fact the condition N=
= 0 is the required terminating condition.
Note: Since recursion involves overheads such as CPU time and
memory storage, it is suggested that recursion should be used with care.
Example 8: Write a function that computes xy by using recursion.
Solution: We can use the property that xy is simply a product of x and xy1
. For example, 64 = 6 63 . The recursive definition of xy is given below:

The required function is given below :

//Functiontocomputex^y
intpower(intx,inty)
{
if(y==0)
return(1);
else
return(x*power(x,y1));
}
Note: The character ^ has been used to indicate power operator.

Example 9: Write a program that computes GCD (greatest common

divisor) of given two numbers.
Solution: The greatest common divisor of two numbers can be
computed by the algorithm given below:
Algorithmgcd()

Steps
1. Find the larger of the two numbers and store larger in x and smaller in y.
2. Divide the x by y and store the remainder in rem.
3. If rem is equal to zero, then the smaller number (y) is the required GCD and
stop.

else store y in x and rem in y and repeat steps 2 and 3

The above given algorithm can be used to develop a recursive function
called GCD which takes two arguments x and y and integer type and
returns the required greatest common divisor.
//ThisfunctioncomuptesGCD
intgcd(intx,inty)
{
intrem;
rem=x%y;
if(rem==0)
returny;
else
gcd(y,rem);
}
Therequiredprogramthatusesthefunctiongcd()is
givenbelow:
/*ThisprogramcomputesGCDoftwonumberswiththe
helpofafunctiongcd().*/
#include<stdio.h>
intgcd(int,int);
voidmain()
{
intx,y,ans;

printf("\nEntertheintegerswhosegcdistobefound
:");
scanf("%d%d",&x,&y);
if(x>y)
ans=gcd(x,y);
else
ans=gcd(y,x);
printf("\nTheGCDis:%d",ans);
}
//ThisfunctioncomputestheGCD
intgcd(intx,inty)
{
intrem;
rem=x%y;
if(rem==0)
returny;
else
gcd(y,rem);
}
Example 10: Write a program that generates the first n terms of
Fibonacci sequence by recursion. The sequence is 0, 1, 1, 2, 3, 5, 8, .
Solution: In a fibonacci series each term (except the first two) can be
obtained by the sum of its two immediate predecessors. The recursive
definition of this sequence is given below:

Let us now write a program that uses a function Fib() to compute the
first n terms of the series.
/*ThisprogramgeneratesFibonacciseries*/
#include<stdio.h>
intfib(int);
voidmain()
{
intn;
inti,term;

printf("\nEnterthetermstobegenerated:");
scanf("%d",&n);
for(i=1;i<=n;i++)
{
term=fib(i);
printf("%d",term);
}
}
//Functiontoreturnafibonacciterm
intfib(intn)
{
if(n==1)
return0;
else
if(n==2)
return1;
else
return(fib(n1)+fib(n2));
}
From above, we can summarize the characteristics of recursion as:
o There must exist at least a basic solution called terminating condition. This
statement is processed without recursion.
o There must exist a method of reducing a problem into a smaller version of
the problem
o A mechanism to throw back the smaller problem back to the recursive
function

1.17.1 Types of Recursion

Depending upon the way the recursive function is called, the recursion
process can be divided into two categories: direct recursion and mutual
recursion. In direct recursion the recursive function is called from within
itself. In mutual recursion, two recursive functions are involved and they
mutually call each other as shown in Figure 1.12.
The mutual recursion is also called as indirect recursion. It may be noted
that in direct recursion the recursive function myFunc() is calling itself,
whereas in mutual recursion the two recursive functions: myFunc() and
yourFunc() are mutually calling each other.

Fig. 1.12 Types of recursion

The direct recursion can be further divided into three categorieslinear,
binary and tail recursion. Though there are many more terms in vogue
such as tree recursion etc, but we will discuss the popular forms of
recursion as given below:
A. Linear Recursion
When a recursive function has a simple repetitive structure and calls
itself once then it is called as linear recursion. For example, the recursive
function called power(), given below, is a linear recursive function.
intpower(intx,inty)
{
if(y==0)
return(1);
else
return(x*power(x,y1));
}
It may be noted that it checks the terminating condition and thereafter
performs the single recursive call to itself.
B. Binary Recursion
When a recursive function calls itself twice from within the function then
it is called as binary recursion. For example, the fib(), given below, calls
itself twice i.e. for fib(n 1) and fib(n 2).
//Functiontoreturnafibonacciterm
intfib(intn)

{
if(n==1)
return0;
else
if(n==2)
return1;
else
return(fib(n1)+fib(n2));
}
Another example of binary recursion is binary tree recursive operations,
done for both left child and right child sub trees.
C. Tail Recursion
When a recursive call in a recursive function is the last call without any
pending operation then the recursion is called tail recursion. Thus, the
last result of the recursive call is the final result of the function.
Consider the function fact(), given below:
longintfact(intn)
{
if(n==0)
return(1);
else
return(n*fact(n1));
}
This function is not tail recursive because the last statement is not
recursive call but the multiplication operation: n * fact (n 1).
However, the following modified function fact() is tail recursive.
longfact(intN,intresult)
{
if(N==1)
returnresult;
else
returnfact(N1,N*result);//Tailrecursion
}
It may be noted that the last statement in the above function is a
recursive call without any pending operation. Hence the function is tail
recursive. Since the argument &result& is acting as an accumulator, its
initial value must be set to 1. Therefore, for computing factorial of a

given number (say 5), the above function must be called with the
following statement:
Factorial = fact ( 5, 1);
Example 11: Write a program that computes the factorial of a given
number by using the above given tail recursive function fact().
Solution: The required program is given below:
#include<stdio.h>
longfact(intN,intresult)
{
if(N==1)
returnresult;
else
returnfact(N1,N*result);
}
voidmain()
{
longfactorial;
intnum;
printf("\nEnterthenumber");
scanf("%d",&num);
factorial=fact(num,1);
printf("\nfactorial=%u",factorial);
}
Sample outputs for the program are given below:

Example 12: Write a tail recursive function for computing xy.

Solution: The required function power (x,y) is given below:

longpower(intx,inty,longresult)
{
if(y==0)
returnresult;

else
returnpower(x,y1,x*result);//Tail
recursion
}
Example 13: Write a program that tests the tail recursive function
of Example 12 for given values of x and y.
Solution: The required program is given below:
#include<stdio.h>
longpower(intx,inty,longresult)
{
if(y==0)
returnresult;
else
returnpower(x,y1,x*result);//Tailrecursion
}
voidmain()
{
longpow;
intx,y;
printf("\nEnterthex,y");
scanf("%d%d",&x,&y);
pow=power(x,y,1);
printf("\nx^y=%u",pow);
}
Sample output of the above program is given below:

Example 14: Write both normal and tail recursive versions of a

function sum(N) that adds the first N integers. For example, sum(6) is
computed as following:

Sum(6) = 1 + 2 + 3 + 4 + = + 6
Solution: The normal recursive function is given below:

longsum(intN)
{
if(N==1)
return1;
else
returnN+sum(N1);
}
The tail recursive version of function sum() is given below:
longsum(intN,intresult)
{
if(N==1)
returnresult;
else
return(sum(N1,N+result));//Tailrecursion
}
Example 15: Write a complete program that takes the tail recursive
version of the function sum() to compute the first N integers.

Solution: The required program is given below:

#include<stdio.h>
longsum(intN,intresult)
{
if(N==1)
returnresult;
else
return(sum(N1,N+result));
}
voidmain()
{
intN,total;
printf("\nEnterN");
scanf("%d",&N);
total=sum(N,1);
printf("\nsum=%d",total);
}

Sample output of the program is given below:

1.17.2 Tower of Hanoi

The problem of Tower of Hanoi, based on a legend of Vietnam, was first
described by the mathematician Franois duoard Anatole Lucas. Later
on Henri de Parville also described the same problem based on an
ancient Indian legend. It says that the dome of the Kashi Vishwanth
Temple at Banaras marks the centre of the world. Under the dome there
is Tower of Bramah.
The statement of the problem is that there are three needles like towers.
On one tower there are 64 disks of decreasing sizes. The largest disk
being at the bottom of the tower and the smallest at the top as shown
in Figure 1.13.

Fig. 1.13 The Tower of Hanoi

A group of monks or priests are moving the disks from source tower to
destination tower as per the following rules:
1. At the beginning of the creation, the disks were placed in decreasing size on
the source tower
2. The mandate is to move the disks to one of the other tower (say destination)
with the help of a third tower called middle.
3. Only one disk can be moved at a time, taken from top of the tower
4. A disk can never be placed on a smaller disk.

The legend predicts that as soon as all the disks are moved to the
destination tower, the world will come to an end, i.e., the priests will
crumble into dust and the world will vanish.
This problem can be solved with the help of recursion. Assume there are
N disks on the source tower. Since on the destination tower, the
bottommost disk needs to be the bottom most of the source tower, it is
necessary that we move N 1 disks to the middle tower and thereafter

move the largest (bottommost) disk to the destination tower as shown

in Figure 1.14.

Fig. 1.14 Moving N 1 disks to the middle tower and the bottom most
to final tower
Now repeat this exercise with N 1 disks using source tower as the
middle tower, and middle tower playing the role of the source tower.
The above given steps can be written using a recursive algorithm as given
below:
Algoritm towOfHan ( N, source, dest, middle )
{
if(N==1)
moveadiskfromsourcetodest
else
{towOfHan(N1,source,middle,dest)
moveadiskfromsourcetodest
towOfHan(N1,middle,dest,source)
}
}
Example 16: Write a program that simulate the moves of Tower of
Hanoi for a given number of disks, say N.
Solution: The above given algorithm towOfHan() has been used to code
the required program which is given below:
//ThisprogramsimulatesthemovesoftowerofHanoi
#include<stdio.h>
#include<conio.h>
voidmovDisk(intsource,intdest);
voidtowOfHan(intnumDisk,intsource,intdest,int
mid);
voidmain()
{

intnumDisk;
clrscr();
printf("\nHowmanydisksdoyouwanttomove?");
scanf("%d",&numDisk);
printf("\nDiskTomove%ddisksfrom1to2using3
asmiddledisk:",
numDisk);
towOfHan(numDisk,1,2,3);
}
voidmovDisk(intsource,intdest)
{
printf("\nmoveadiskfrom%dto%d",source,dest);
}
voidtowOfHan(intnumDisk,intsource,intdest,int
mid)
{
if(numDisk==1)
movDisk(source,dest);
else
{
towOfHan(numDisk1,source,mid,dest);
movDisk(source,dest);
towOfHan(numDisk1,mid,dest,source);
}
}
A sample output of the program is given below:

TEST YOUR SKILLS

1. What would be the output of following program?
#include<stdio.h>
main()
{

printf(Compu);
printf(ter);
}
Ans. The output would be: Computer
2. What would be the output of following program?
#include<stdio.h>
main()
{
printf(Compu\buter);
}
Ans. The output would be: Computer
3. What would be the output of following program?
voidmain()
{
intk=30;
/*
printf(k=%d,k);
*/
printf(valueofk=%d,k);
}
Ans. The output would be: value of k = 30
4. Write four equivalent C statements each of which subtracts 1 from a variable
called Val.
Ans. The required statements are:

1.
2.
3.
4.

Val=Val1;
Val;
Val;
Val=1;

2 Write the following expressions without using relational operators:

1.
2.

if(Val==0)
if(Val!=0)

Ans. Since in C 1 represents true and 0 represents false, the following

statements can be used instead of above:

1.
2.

if(!Val)
if(Val)

2 Write a single statement that increments the value of a variable Val by 1 and
then adds it to a variable called num.
Ans. The required statement is: num=num+(++Val);
3 What will be the output of the following statements?

1.
2.

ch=A+10;
printf(%c,ch);
inta=5;
printf(%d,a<10);

4 Ans. The output would be: (i) K, (ii) 1

2 What would be the output of following program?
#include<stdio.h>
main()
{
intk=10;
k=++k==11;
printf(\n%d,k);
}
Ans. The output would be: 1
3 What would be the output of following program?
#include<stdio.h>
main()
{
inti=8>5?100:150;
printf(\n%d,i);
}
Ans. The output would be: 100
4 What would be the output of the following program?
enumfurniture{Chair,
Table,
Bed
Almirah
};
#include<stdio.h>
main()
{
printf(\n%d%d%d%d,Chair,Table,Bed,Almirah);
}

Ans. The output would be: 0 1 2 3

5 What would be the output of the following program?

#include<stdio.h>
intfirst(int,int);
intsecond(int);
voidmain()
{
intx=10,y=5,z;
z=first(x,y);
printf(\nz=%d,z);
}
intfirst(intx,inty)
{
return(x+second(y));
}
intsecond(inty)
{
return++y;
}
Ans. The output would be: 16

EXERCISES
1. Define the terms: token, keyword, identifier, variable, constant and const
correct.
2. What is meant by basic data types? Explain in brief.
3. What would be the appropriate data type for the following?

1. Length of a cloth
2. Age of a student
3. An exclamation mark
2 What is meant by a backslash character? What is the utility of these
characters in a program?
3 What would be the output of the following program?
#include<stdio.h>
{
intk;
printf(k=%d,k);
}

1.
2.
3.
4.

an error
unpredictable
34216
0

2 Write a program that inputs a value in inches and prints in centimeters.

3 Write a program that converts degree Celsius temperature into its degree
Fahrenheit equivalent using the following formula:
F = 9/5 C + 32
4 What is the purpose of else clause in an if statement of C?
5 Write a program that computes ab where a and b are of real and integer
types, respectively.
6 Describe the function of break and continue statements in C.
7 Write a program that reverses and sums the digits of an integer number.
8 Write a program that prints the following output on the screen.
A
BB
CCC
DDDD
EEEEE
9 Write a program that implements a menu of the following type:

In response to users selection, it prints the following messages:

10 Define the terms: arrays, subscript, subscripted variables, and strings.

11 Write a program that removes duplicates from a list.
12 Write a program that computes the sum of diagonal elements of a square
matrix.
13 Write a C structure for the record structure given in figure shown below.

14 Define the terms: structure and member variables.

15 Using the structure of following figure, write a program that maintains a list
of computer books available in the library.
16 What is meant by enumerated data types? Explain the concept with the help
of suitable examples.
17 What is meant by a function prototype and why is it needed?
18 Define recursion and compare it with iteration.
19 What are the types of recursion?
20 Explain tail recursion with the help of a suitable example.
21 What are the basic requirements of a recursive algorithm?
22 Write a program that uses recursion to add first N natural numbers
23 Explain the problem of Tower of Hanoi. Write a recursive function that
moves N disks from a tower called first to tower last using a tower mid.

Note: Kindly refer to Chapter 5 before reading the text of this section.

2
Data Structures and Algorithms: An Introduction
CHAPTER OUTLINE

2.1 Overview
2.2 Concept of Data Structures
2.3 Design of a Suitable Algorithm
2.4 Algorithm Analysis
2.1 OVERVIEW

A normal desktop computer is based on Von Neumann architecture. It is

also called a stored program computer because the programs
instructions and its associated data are stored in the same memory as
shown in Figure 2.1.
The program instructions tell the system what operation is to be
performed and on what data.
The above model is based on the fact that every practical problem that
needs to be solved must have some associated data. Let us consider a
situation where a list of numbers is to be sorted in ascending order. This
problem obviously has the following two parts:
1. The list: say 9, 8, 1, -1, 4, 2, 6, 15, 3
2. A procedure to sort the given list of numbers.

It may be noted that elements of list and the procedure represent the
data and program of the problem, respectively.
Now, the efficiency of the program would primarily depend upon the
organization of data. When the data is poorly organized, then the
program cannot efficiently access it. For instance, the time taken to open
the lock of your house depends upon where you have kept the key to it:
in your hand, in front pocket, in the ticket pocket of your pant or the
inner pocket of your briefcase, etc. There may be a variety of motives for
selecting one or the other of these places to keep the key. Key in the hand
or in the front pocket would be quickly accessible as compared to other

places whereas the key placed in the inner pocket of briefcase is

comparatively much secured. For similar reasons, it is also necessary to
choose a suitable representation or structure for data so that a program
can efficiently access it.

Fig. 2.1 The stored program computer

2.2 CONCEPT OF DATA STRUCTURES

In fact the data, belonging to a problem, must be organized for the

following basic reasons:
1. Suitable representation: Data should be stored in such a manner that it
represents the various components of the problem.
For example, the list of numbers can be stored in physical continuous
locations. Arrays in C can be used for the required representation as shown
in Figure 2.2.

Fig. 2.2 List stored in physical continuous locations

2. Ease of retrieval: The data should be stored in such a manner that the
program can easily access it.
As an array is an indexvalue pair, a data item stored in the array
(see Figure 2.2) can be very easily accessed through its index. For
example, List[5] will provide the data item stored at 5th location, i.e. 2.
From this point of view, the array is a random-access structure.
3. Operations allowed: The operations needed to be performed on the data
must be allowed by the representation.
For example, component by component operations can be done on the data
items stored in the array. Consider the data stored in List of Figure 2.2.
For the purpose of sorting, the smallest number must come to the head of
the list. This shall require the following two operations:

1. Visit every element (component) in the list to search the

smallest and its position (-1 at location 3).
2. Exchange the smallest with the number stored at the
head of the list (exchange 1 stored at location 3 with 9
stored at location 0). This can be easily done by using a
temporary location (say Temp) as shown below:
Temp=List[0];
List[0]=List[3];
List[3]=Temp;
The new contents of the List are as shown in Figure 2.3.

Fig. 2.3 Contents of the List after exchange operation

The operations 1 and 2, given above, can be repeated on the unsorted
part of the List to obtain the required sorted List.
From above discussions, it can be concluded that an array provides a
suitable representation for the storage of a list of numbers. Such a
representation is also called a data structure. Data structure is a
logical organization of a set of data items that collectively describe an
object. It can be manipulated by a program. For example, array is a
suitable data structure for a list of data items.
It may be observed that if the array called List represents the data
structure then the operations carried out in steps 1 and 2 define
the algorithm needed to solve the problem. The overall performance of a
program depends upon the following two factors:
1. Choice of right data structures for the given problem.
2. Design of a suitable algorithm to work upon the chosen data structures.

2.2.1 Choice of Right Data Structures

An important step of problem solving is to select the data structure
wherein the data will be stored. The program or algorithm would then
work on this data structure and give the results.
A data structure can either be selected from the built-in data structures
offered by the language or by designing a new one. Built-in data
structures in C are arrays, structures, unions, files, pointers, etc.
Let us consider a situation where it is required to create a merit list of 60
students. These students appeared in a class examination. The data of a
student consists of his name, roll number, marks,percentage,

and grade. Out of many possibilities, the following two choices of data
structures are worth considering:
Choice I: The various data items related to a student are: name, roll
number, marks, percentage, and grade. The total number of students is
60. Therefore, the simplest choice would be to have as many parallel lists
as there are data items to be represented. The arrangement is
shown Figure 2.4.

Fig. 2.4 Parallel lists for student data

An equivalent declaration for data structure of Figure 2.4 is given below:
charnameList
[60][15];
introll[60];
intmarks[60];
floatpercent[60];
chargrade[60];
Now in the above declaration, every list has to be considered as
individual. The generation of merit list would require the list to be
sorted. The sorting process would involve a large number of exchange
operations. For each list, separate exchange operations would have to be
written. For example, the exchange between data of ith and jth student
would require the following operations:
chartempName[15],chTemp;
inti,j,temp;
floatfTemp;
strcpy(tempName,nameList[i]);
strcpy(nameList[i],nameList[j]);
strcpy(nameList[j],tempName);

temp=roll[i];
roll[i]=roll[j];
roll[j]=temp;
temp=marks[i];
marks[i]=marks[j];
marks[j]=temp;
fTemp=percent[i];
percent[i]=percent[j];
percent[j]=fTemp;
chTemp=grade[i];
grade[i]=grade[j];
grade[j]=chTemp;
Though the above code may work, but it is not only a clumsy way of
representing the data but would require a lot of typing effort also. In fact,
as many as fifteen statements are required to exchange the data of ith
student with that of jth student.
Choice II: From the built-in data structures, the struct can be
selected to represent the data of a student as shown below:
structstudent{
charname[15];
introll;
intmarks;
floatpercent;
chargrade;
};
Similarly, from built-in data structures, an array called studList can
be selected to represent a list. However, in this case each location has to
be of type student which itself is a structure. The required declaration is
given below:
structstudentstudList[60];
It may be noted that the above given array of structures is a user-defined
data structure constructed from two built-in data
structures: struct and array. Now in this case, exchange between data
of ith andjth student would require the following operations:
inti,j;

structstudenttemp;
temp=studList[i];
studList[i]=studList[j];
studList[j]=temp;
Thus, only three statements are required to exchange the data of ith
student with that of jth student.
A comparison of the two choices discussed above establishes that Choice
II is definitely better thanChoice I as far as the selection of data
structures for the given problem is concerned. The number of lines of
code of Choice II is definitely less than the Choice I. Moreover, the code
of Choice II is comparatively easy and elegant.
From the above discussion, it can be concluded that choice of right data
structures is of paramount importance from the point of view of
representing the data belonging to a problem.
2.2.2 Types of Data Structures
A large number of data structure are available to programmers that can
be usefully employed in a program. The data structures can be broadly
classified into two categories: linear and non-linear.
(1) Linear data structures: These data structures represent a
sequence of items. The elements follow a linear ordering. For instance,
one-dimensional arrays and linked lists are linear data structures. These
can be used to model objects like queues, stacks, chains (linked lists),
etc., as shown in Figure 2.5.

Fig. 2.5 Linear data structures

A linear data structure has following two properties:

1. Each element is followed by at most one other element.

2. No two elements are followed by the same element.
For example, in the list of Figure 2.3, the 2nd element (1) is followed by 3rd
element (9) and so on.

(2) Non-linear data structures: These data structures represent

objects which are not in sequence but are distributed in a plane. For
instance, two-dimensional arrays, graphs, trees, etc. are non-linear data
structures. These can be used to model objects like tables, networks and
hierarchy, as shown inFigure 2.6.

Fig. 2.6 Non-linear data structures

The tree structure does not follow the first property of linear data
structures and the graph does not follow any property of the linear data
structures. A detailed discussion on this issue is given later in the book.
2.2.3 Basic Terminology Related with Data Structures
The basic terms connected with the data structures are:
1. Data element: It is the most fundamental level of data. For example,
a Date is made up of three parts: Day, Month, and Year as shown in Figure
2.7. The Date is called a data object and DD, MM, and YY are called the data
elements or data items.
It may be noted that the data object denotes a group and the data elements
denote atomic entities.
2. Primitive data types: It is type of data that a variable may hold. For
example, int, float, char, etc. are data types.
3. Data object: It is a set of elements. For example, the set of fruit F refers
to the set F = {mango, orange, grapes}. Similarly, a set of counting
numbers C refers to the set C = {1, 2, 3, 4.}.

Fig. 2.7 Data elements of data object Date

4. Data structure: It is a set of related elements.A set of operations that are
allowed on the elements of the data structure is defined.

5. Constant: It is the memory location that does not change its contents
during the execution of a program.
6. Variable: It is the most fundamental aspect of any computer language and
can be defined as a location in the memory wherein a value can be
manipulated, i.e., stored, accessed, and modified.
2.3 DESIGN OF A SUITABLE ALGORITHM

After selecting the data structures, an algorithm, based on the data

structure, needs to be designed. The set of rules that define how a
particular problem can be solved in a finite sequence of steps is called
analgorithm. An algorithm written in a computer language is called
a program. An algorithm can be defined as a finite sequence of
instructions, each of which has a clear meaning and can be executed with
a finite amount of effort in finite time.
A computer is simply a machine with no intelligence. In fact, the
processor of the computer is a passive device that can only execute
instructions. Therefore, the computer must be instructed in
unambiguous steps such that it follows them one after the other to
perform the given task. Thus, it becomes programmers duty to precisely
define his problem in the form of an algorithm, i.e., unambiguous steps
written in simple English. The desirable features of an algorithm are:
1. Each step of the algorithm should be simple.
2. It should be unambiguous in the sense that the logic should
be crisp and clear.
3. It should be effective, i.e., it must lead to a unique solution of the problem.
4. It must end in a finite number of steps.
5. It should be as efficient as possible.

Accordingly, an algorithm has the following characteristics:

1. Input: This part of the algorithm reads the data of the given problem.
2. Process: This part of the algorithm does the required computations in
simple steps.
3. Finiteness: The algorithm must come to an end after a finite number of
steps.
4. Effectiveness: Every step of the algorithm must be accurate and precise. It
should also be executable within a definite period of time on the target
machine.
5. Output: It must produce the desired output.

Thus, we can say that the algorithm tells the computer when to read
data, compute on it, and write the answers.

2.3.1 How to Develop an Algorithm?

In order to develop an algorithm, the problem has to be properly
analysed. The programmer has to understand the problem and come out
with a method or logic that can solve the problem. The logic must be
simple and adaptable on a computer. The following steps can be
suggested for developing an algorithm:
1. Understand the problem.
2. Identify the output of the problem.
3. Identify inputs required by the problem and choose the associated data
structures.
4. Design a logic that will produce the desired output from the given inputs.
5. Test the algorithm for different sets of input data.
6. Repeat steps 1 to 5 till the algorithm produces the desired results for all
types of input and rules.

While solving problems with the help of computers, the following facts
should be kept in mind:
A location in the main memory has the property: Writing-in is
destructive and Reading-out is non-destructive. In simple words, we
can say that if a value is assigned to a memory location, then its previous
contents will be lost. However, if we use the value of a memory location
then the contents of the memory do not change, i.e., it remains intact.
For example, in the following statement, the value of variables B and C
are being used and their sum is stored in variable A
A=B+C
where, the symbol = is an assignment operator.
If previous values of A, B, C were 15, 45, 20, respectively then after
execution of the above statement, the new contents of A, B, C would be
65, 45, and 20, respectively.
Consider the following statement:
Total = Total + 100;
The abnormal looking statement is very much normal as far as computer
memory is concerned. The statement reveals the following fact:
Old value of variable Total is being added to an integer 100 and the
result is being stored as new value in the variable Total, i.e., replacing
its old contents with the new contents.
After the execution of the statement with old contents of Total (say 24),
the new contents of Total will be 124.

2.3.1.1 Identifying Inputs and Outputs

It is very important to identify inputs and outputs of an algorithm. The
inputs to an algorithm mean the data to be provided to the target
program through input devices such as keyboard, mouse, input file, etc.
The output of the algorithm means the final data to be generated for the
output devices such as monitor, printer, output file, etc.
The inputs and outputs have to be identified at the beginning of the
program development. The most important decision is to differentiate
between major and minor inputs and outputs. The minor inputs are
sometimes omitted at the algorithm development stage. The other
important issue is to find at which part of the program the inputs would
be required and the output generated.
Example 1: Let us reconsider the problem of sorting a given list of
numbers. In this case, the requiredinputs are:
1.
2.
3.
4.

List of numbers
Size of list
Type of numbers to be stored in the list
In which order to be sorted

The outputs are:

1. The sorted list
2. Message for asking the size of the list
3. Message for displaying the output list

The messages are the minor outputs.

After identifying the inputs and the outputs, the additional information
must be specified before writing the final algorithm.
Input specifications
1. In what order and format, the input values will be read?
2. What are the lower and upper limits of the input values? For example, the
size of the list should not be less than zero and its type has to be necessarily
an integer.
3. When to know that there is no more input to be read, i.e., what marks the
end of the list?

Output specifications
1. In what order and format the outputs values will be produced?
2. What types of values to be produced with what spacing i.e., decimal spacing,
significant digits, etc.
3. What main headings and column headings to be printed on the output page.

2.3.2 Stepwise Refinement

We know that if there is even a small logical error in a program, the
compiler cannot detect it. The year 2000 problem (Y2K problem) is a
glaring example of this type. Thus, an algorithm requires a very careful
design, especially pertaining to a large and complex problem.
In order to design an algorithm, the programmer must adopt some
methodological approach such as modular/stepwise refinement
approach of structured programming techniques. Stepwise refinement
uses the divide and conquer strategy. In this technique, the problem is
divided into a number of steps. Each step is described by a sub-problem
which is smaller and simpler as compared to the original program. The
sub-problem is again divided into still smaller steps and so on. The
refinement continues till we reach to a stage where the steps become
fairly simple to implement.
Example 2: Suppose the following table is to be printed on the screen.

The problem at hand is How to print the table. The first version of the
algorithm is given below:
Algorithm Print_TableVer1.
{
Step1.Printtable
}
When we look at the table, we find that the table is nothing but a
collection of rows and we have to print only 10 rows. So, the problem has
been reduced to printing a row 10 times. The second refined version of
the algorithm is given below:
Algorithm Print_TableVer2.
{
Step
1.Row=1
2.Printtherow
3.Row=Row+1
4.IfRow<=10repeatsteps2and3
5.Stop
}

Now our focus is on how to print a row. A closer look at the problem
suggests that a row is nothing but a collection of columns and we have to
print only 10 columns for each row. The third refined version of the
algorithm is given below:
Algorithm Print_TableVer3.
{
Step
1.Row=1
2.Col=1
2.1Printthecol.
2.2Col=Col+1
2.3IfCol<=10repeatsteps2.1and2.2
2.4Gotoanewline.
3.Row=Row+1
4.IfRow<=10repeatsteps2and3
5.Stop.
}
Now our concern is how to print a column. We find that it is very
simple, i.e., we get the value to be printed by multiplying the values
contained in the variables Row and Col The final version of the refined
algorithm is given below:
Algorithm Print_TableVer4.
{
Step
1.Row=1
2.Col=1
2.1ColVal=Row*Col
2.2PrinttheColVal.
2.3Col=Col+1
2.4IfCol<=10repeatsteps2.1to2.3
2.5Gotoanewline.
3.Row=Row+1
4.IfRow<=10repeatsteps2and3
5.Stop.
}
The programmer stops at this stage where he finds that the step(s) is
trivial and can be easily implemented.

2.3.3 Using Control Structures

Every language supports sequence control structures. These structures
can be easily used towards writing a program for a given problem.
(1) Sequence structure: The order in which the statements or
statement blocks are written within a pair of curly braces ({,}) in an
algorithm are sequentially executed by the computer.
Example 3: Write an algorithm that computes the area of a triangle.
Solution: The solution is trivial. The inputs are base and height. The
following formula will be used for the computation of the area of
triangle:
Area = * base * height
where area is the output to be displayed. The simple sequence structure
required for this problem is given below:
Algorithm AreaTriangle()
{
Step
1.Readbase,height;
2.Area=0.5*base*height;
3.PrintArea;
4.Stop
}
It may be noted that in a sequence structure:
1.
2.
3.
4.
5.

The statements are executed sequentially.

The statements are executed one at a time.
Each statement is executed only once.
No statement is repeated.
The execution of the last statement marks the end of the sequential
structure.

(2) Selection (conditional control) structure: The selection or

decision structure is used to decide whether a statement or a block is to
be executed or not. The decision is made by testing whether a condition
is true or false. The if-else ladder structure can be used for this purpose.
Example 4: Write an algorithm that prints the biggest of given three
unique numbers.

Solution:
Input: Three numbers: Num1,Num2, and Num3
Type of numbers: integer.
Output: Display the biggest number and a message in this regard.
Algorithm Biggest()
{
Step
1.ReadNum1,Num2,Num3;
2.If(Num1>Num2)
2.1If(Num1>Num3)
PrintNum1isthelargest
Else
PrintNum3isthelargest
Else
2.2If(Num2>Num3)
PrintNum2isthelargest
Else
PrintNum3isthelargest
3.Stop
}
(3) Iteration: The iteration structure is used to repeatedly execute a
statement or a block of statements. The loop is repeatedly executed until
a certain specified condition is satisfied.
Example 5: Write an algorithm that finds the largest number from a
given list of numbers.
Solution:
Input: List of numbers, size of list, prompt for input.
Type of numbers: integer.
Output: Display the largest number and a message in this regard.
The most basic solution to this problem is that if there is only one
number in the list then the number is the largest. Therefore, the first
number of the list would be taken as largest. The largest would then be
compared with the rest of the numbers in the list. If a number is found
larger than the current largest, then it is stored in the largest. This
process is repeated till the end of the list. For iteration, For loop can be
employed.
The concerned algorithm is given below:

Algorithm Largest
{
Step
1.PromptEnterthesizeofthelist;
2.ReadN;
3.PromptEnteraNumber;
4.ReadNum;
5.Largest=Num;/*Thefirstnumberisthe
largest*/
6.For(I=2toN)
{
6.1PromptEnteraNumber
6.2ReadNum
6.3If(Num>Largest)Largest=Num
}
7.PromptThelargestNumber=
8.PrintLargest;
9.Stop
}
It may be noted here that if a loop terminates after a number of
iterations, then the loop is called afinite loop. On the other hand if a
loop does not terminate at all and keeps on executing endlessly, then it is
called an infinite loop.
(4) Use of accumulators and counters: Accumulator is a variable
that is generally used to compute sum or product of a series. For
example, the sum of the following list of numbers can be computed by
initializing a variable (say Total) to zero.
The 1st number of the list is added to Total and then the 2nd, 3rd and so
on till the end of the list. Finally, the variable Total would get the
accumulated sum of the list given above.
Example 6: Write an algorithm that reads a list of 50 integer values and
prints the total of the list.
Solution: Let us assume that a variable called sum would be used as an
accumulator. A variable called Num would be used to read a number
from the input list one at time. For loop will be employed to do the
iteration.
The Algorithm is given below:

Algorithm SumList
{
Step
1.Sum=0;
2.For(I=1to50)
{
2.1ReadNum;
2.2Sum=Sum+Num;/*Theaccumulationofthe
list*/
}
3.PromptThesumofthelist=;
4.PrintSum;
5.Stop
}
It may be noted that for product of a series, the accumulator is initialized
to 1.
Counter is a variable used to count number of operations/actions
performed in a loop. It may also be used to count number of members
present in a list or a set of items.
The counter variable is generally initialized by 0. On the occurrence of an
event/operation, the variable is incremented by 1. At the end of the
counting process, the variable contains the required count.
Example 7: Write an algorithm that finds the number of zeros present
in a given list of N numbers.
Solution: A counter variable called Num_of_Zeros would be used to
count the number of zeros present in a given list of numbers. A variable
called Num would be used to read a number from the input list one at
time. For loop will be employed to do the iteration.
The Algorithm is given below:
Algorithm Count_zeros()
{
Step
1.Num_of_zeros=0;
2.For(I=1toN)
{
2.1ReadNum;
2.2If(Num==0)
Num_of_zeros=Num_of_zeros+1;
}

3.PromptThenumberofzeroelements=;
4.PrintNum_of_zeros;
5.Stop
}
Example 8: Write a program that computes the sum of first N multiples
of an integer called K, i.e.
sumSeries = 1*K + 2*K + 3*K + N*K
Solution: An accumulator called sumSeries would be used to
compute the required sum of the series. As the number of steps are
known (i.e., N), For loop is the most appropriate for the required
computation. The index of the loop will act as the counter.
Input: N, K
Output: sumSeries
/*Thisprogramcomputesthesumofaseries*/
#include<stdio.h>
voidmain()
{
intsumSeries,i;
intN;/*Numberofterms*/
intK;/*Theintegermultiple*/
printf(\nEnterNumberoftermsandInteger
Multiple);
scanf(%d%d,&N,&K);
sumSeries=0;/*Accumulatorinitialized*/
For(i=1;i<=N;i++)
{
sumSeries=sumSeries+i*K;
}
printf(\nThesumoffirst%dterms=%d,N,
sumSeries);
}
In fact, it can be concluded that irrespective of the programming
paradigms, a general algorithm is designed based on following
constructs:
1. Sequence structure: Statement after statements are written in sequence
in a program.

2. Selection: Based on a decision (i.e., if-else), a section of statements is

executed.
3. Iteration: A group of statements is executed within a loop.
4. Function call: At a certain point in program, the statement may call a
function to perform a particular task.

It is the responsibility of the programmer to use the right construct at

right place in the algorithm so that an efficient program is produced. A
discussion on analysis of algorithms is given in next section.
2.4 ALGORITHM ANALYSIS

A critical look at the above discussion indicates that the data structures
and algorithms are related to each other and have to be studied together.
For a given problem, an algorithm is developed to solve it but the
algorithm needs to operate on data. Unless the data is stored in a
suitable data structure, the design of algorithm remains incomplete. So,
both the data structures and algorithm are important for program
development.
In fact, an algorithm requires following two resources:
1. Memory space: Space occupied by program code and the associated data
structures.
2. CPU time: Time spent by the algorithm to solve the problem.

Therefore, an algorithm can be analyzed on two

accounts: space and time We need to develop an algorithm that
efficiently makes use of these resources. But then, there are many ways
to attack a problem and hence many algorithms at hand to choose from.
Obviously, having decided the data structures, execution time can be a
major criterion for selection of an algorithm.
As a first step, it is understandable that an algorithm would be fast for a
small input data. For example, the search for an item in a list of 10
elements would definitely be faster as compared to searching the same
item within a list of 1,000 elements. Therefore, the execution time T of
an algorithm has to be dependent upon N, the size of input data. Now the
question is how to measure the execution time T. Can we exactly
compute the execution time? Perhaps not because there are many factors
that govern the actual execution of program in a computer system. Some
of the important factors are listed below:
1.
2.
3.
4.

The speed of the computer system

The structure of the program
Quality of the compiler that has compiled the program
The current load on the computer system.

Thus, it is better that we compare the algorithms based on their relative

time instead of actual execution time.
The amount of time taken for completion of an algorithm is called time
complexity of the algorithm. It is a theoretical estimate that measures
the growth rate of the algorithm T(n) for large number of n, the input
size. The time complexity is measured in terms of Big-Oh notation. A
brief discussion on this notation is given in next section.
2.4.1 Big-Oh Notation
The Big-Oh notation defines that for a large number of input n, the
growth rate of an algorithm T(n) is of the order of a function g of n as
indicated by the following relation:
T(n) = O(g(n)).
The term of the order means that T(n) is less than a constant multiple
of g(n) for n >= n0. Therefore, for a constant c, the relation can be
defined as given below:
T(n) <= c*g(n)

where c>0.

(2.1)

From above relation, we can say that for a large value of n, the function
g provides an upper bound on the growth rate T.
Let us now consider the various constructs used by a programmer to
write an algorithm.
(1) Simple statement: Let us assume that a statement takes a unit
time to execute, i.e., 1. Thus, T(n) = 1 for a simple statement. To satisfy
the Big-Oh condition, the above relation can be rewritten as shown
below:
T(n) <= 1*1

where c = 1 and n0= 0.

Comparing the above relation with relation 2.1, we get g(n) = 1.

Therefore, the above relation can be expressed as:
T(n) = O(1)
(2) Sequence structure: The execution time of sequence structure is
equal to the sum of execution time of individual statements present
within the sequence structure.
Consider the following algorithm. It consists of a sequence of statements
1 to 4:
Algorithm AreaTriangle

{
Step
1.Readbase,height;
2.Area=0.5*base*height;
3.PrintArea;
4.Stop
}
The above algorithm takes four steps to complete.
Thus, T(n) = 4.
To satisfy the relation 2.1, the above relation can be rewritten as:
T(n) <= 4*1

where c = 4 and n0 = 0.

Comparing the above relation with relation 2.1, we get g(n) = 1.

Therefore, the above relation can be expressed as:
T(n) = O(1)
We can say that the time complexity of above algorithm is O(1), i.e., of
the order 1.
(3) The loop structure: The loop structure iteratively executes
statements written within its bound. If a loop executes for N iterations, it
contains only simple statements.
Consider the algorithm Count_zeros of Example 7 which contains a
loop. This algorithm takes the following steps for its completion:
No. of simple steps

= 4 (Steps 1, 3, 4, 5)

No. of Loops of steps 1 to N

=1

No. of statements within the

loop

= 3 (1 comparison, 1 addition, 1
assignment)

Thus, T(n) = 3*N + 4

Now for N >= 4, we can say that
3*N + 4 <= 4, 3*N + N <= 4N
Therefore, we can say that
T(n) <= 4N
T(n) = O(N)

where c = 4, n0 >= 4

Hence, the given algorithm has time complexity of the order of N, i.e.,
O(N). This indicates that the term 4 has negligible contribution in the
expression: 3*N + 4.
Note: In case of nested loops, the time complexity depends upon the
counters of both outer and inner loops. Consider the nested loops given
below. This program segment contains S, a sequence of statements
within nested loops I and J.
For(I=1;I<=N;I++)
{
For(J=1;J<=N;J++)
{
S;
}
}
It may be noted that for each iteration of I loop, the J loop executes N
times. Therefore, for N iterations of I loop, the J loop would execute N*N
= N2 times. Accordingly, the statement S would also execute N2times.
Thus, T(n) = N2
To satisfy the relation 2.1, the above relation can be rewritten as shown
below:
T(n) <= 1 * N2

where c = 1 and n0 =0.

Comparing the above relation with relation 2.1, we get g(n) = N2.
Therefore, the above relation can be expressed as:
T(n) = O(N2)
Hence, the given nested loop has time complexity of the order of N2, i.e.,
O(N2).
Example 9: Compute the time complexity of the following nested loop:
For(I=1;I<=N;I++)
{
For(J=1;J<=I;J++)
{
S;
}
}

Solution: In this nested loop structure, for an iteration of I loop, the J

loop executes in an increasing order from 1 to N as given below:
1 + 2 + 3 + 4 (N 1) + N = N*(N + 1)/2 = N2/2 + N/2.
Thus, T(n) = N2/2 + N/2.
Now for N2 <= N/2, we can say that:
N2/2 + N/2 <= N2/2 + N2 <= 3 N2/2.
Therefore, we can say that
T(n) <= 3 N2/2
T(n) = O(N2)

where, c = 3/2, n0 = 1/2

Hence, the given algorithm has time complexity of the order of N2, i.e.,
O(N2). This indicates that the term N/2 has negligible contribution in the
expression N2/2 + N/2.
(4) If-then-else structure: In the if-then-else structure, the then and
else parts are considered separately. Consider the if-then-else structure
given in Figure 2.8.

Fig. 2.8 The if-then-else structure

Let us assume that the statement 1 of then part has the time complexity
T1 and that statement 2 of else part has the time complexity T2. The
time complexity of if-then-else structure is taken to be the maximum of
the two, i.e., max(T1, T2).
Example 10: Consider the following algorithm. Compute its time
complexity.
If(x>y)
{
x=x+1;

}
Else
{
For(i=1;i<=N,i++)
{
x=x+i;
}
}
Solution: It may be noted that in the above algorithm, the time
complexity of then and else part are O(1) and O(N), respectively. The
maximum of the two is O(N). Therefore, the time complexity of the
above algorithm is O(N).
Example 11: compute the time complexity of the following relations:
1.
2.
3.
4.

T(n) = 2527
T(n) = 8*n + 17
T(n) = 15*N2 + 6
T(n) = 5*N3 + N2 = 2*N

Solution:
1. T(n) = 2527 <= 2527*1, where c = 2527 and n0 = 0.
= O(1)
Ans. The time complexity = O(1)

2. T(n) = 8*n + 17 <= 8*n + n <= 9 *n for c = 9 and n0 = 17

= O(N)
Ans. The time complexity = O(N)

3. T(n) = 15*N2 + 6 <= 15*N2 + N for N = 6

Now for N <= N2, we can rewrite the above relation as given below:
15*N2 + N <= 15*N2 + N2 <= 16 N2 for c = 16 and n0 = 6
Thus, T(n) = O(N2)
Ans. The time complexity = O(N2)

4. T(n) = 5*N3 + N2 + 2*N

For N2 >= 2*N, we can rewrite the above relation as given below:
5*N3 + N2 + 2*N <= 5*N3 + N2 + N2 <= 5*N3 + 2*N2
For N3 >= 2*N2, we can rewrite the above relation as given below:

5*N3 + 2*N2 <= 5*N3 + N3 <= 6*N3 for c = 6 and n0 = 2

Thus, T(n) = O(N3)
Ans. The time complexity = O(N3)

Note: All the relations, given in example 10 were polynomials. From the
answers, we can say that any polynomial has the time complexity of BigOh of its leading term with coefficient of 1.
It is also understandable that the Big-Oh does not provide the exact
execution time. Rather, it gives only an asymptotic upper bound of the
execution time of an algorithm.
For different types of functions, the growth rates for different values of
input size n have been tabulated in Table 2.1.
It can be observed from Table 2.1 that for a given input, the growth rate
O(lg n) is faster than O(n lg n). For the same input size, the growth rate
O(N2) is the slowest.
Later in the book, for various algorithms, timing estimates will be
included using Big-Oh notation.
Table 2.1 Growth rates of functions

EXERCISES
1.
2.
3.
4.
5.
6.
7.

8.

What is a stored program computer?

Give the block diagram of Von Neumann architecture of a computer.
What is a data structure? What is the need of data structures?
List out the areas in which data structures are applied extensively. Why is it
important to make a right choice of data structures for problem solving?
What are built-in data structures? Give some examples.
What are user-defined data structures? Give some examples.
Consider a situation where it is required to rank 100 employees of an
organization. The data of an employee consists of his name, designation,
department, experience and salary. Give the possible choices for the data
structures. Draw a comparison between the possible choices of organizing
data.
What is the difference between data type and data structure?

9. Differentiate between linear and non-linear data structures. Explain with

the help of examples.
10. Explain the following basic terminologies associated with the data
structures:

1.
2.
3.
4.
5.
2
3
4
5

6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22

Data element
Primitive data types
Data object
Constant
Variable

What is an algorithm? Differentiate between algorithm and program.

List the desirable features of an algorithm.
What are the characteristics possessed by an algorithm?
What are the steps that can be suggested for developing an algorithm?
Consider a suitable scenario and develop an algorithm for solving the
problem. Explain each step while developing the algorithm.
What are the control structures? Explain the sequence control structures
with the help of examples.
Write an algorithm that computes area and perimeter of a rectangle.
Write an algorithm that computes square of a number by repititive addition.
Write an algorithm that computes GCD and LCM of two given numbers.
Explain the selection (conditional control) structures with the help of
examples.
Write an algorithm that determines whether a number is even or odd.
Write an algorithm that computes profit or loss incurred by a salesperson.
What is iteration?
Write an algorithm that sorts a given list of numbers in ascending order.
Write an algorithm that determines whether a number is prime or not.
What is the use of accumulators and counters?
Write an algorithm that computes average marks of 60 students of a class.
Write an algorithm that finds the number of odds and evens present in a
given list of N numbers.
What are the parameters on the basis of which an algorithm can be
analyzed?
Define the term time complexity. How can the time complexity of a given
algorithm be found?
Explain Big-Oh notation with the help of examples.
Compute the time complexity of the following relations:

1.
2.
3.
4.

T(n) = 410
T(n) = 4*n -12
T(n) = 13*N2 -7
T(n) = 3*N3 + N214*N

2 What is the time complexity of the following loop?

for(inti=0;i<n;i++)
{
statement;
}
3 What is the Big-Oh complexity of the following nested loop?

1.

2.

3.

4.

for(inti=0;i<n;i++)
{
for(intj=0;j<n;j++)
{
statement;
}
}
for(inti=0;i<n;i++)
{
for(intj=i+1;j<n;j++)
{
statement;
}
}
for(inti=0;i<n;i++)
{
for(intj=n;j>i;j)
{
statement;
}
}
for(inti=0;i<n;i++)
{
for(intj=0;j<m;j++)
{
statement1;
}
statement2;
}

2 Consider the following algorithm and compute its time complexity.

if(a==b)

statement1;
elseif(a>b)
statement2;
else
statement3;

Chapter 3
Arrays: Searching and Sorting
CHAPTER OUTLINE

3.1 Introduction
3.2 One-dimensional Arrays
3.3 Multi-dimensional Arrays
3.4 Representation of Arrays in Physical Memory
3.5 Applications of Arrays
3.1 INTRODUCTION

An array is a data structure with the help of which a programmer can

refer to and perform operations on a collection of similar data types such
as simple lists or tables of information. For example, a list of names of
N number of students of a class can be grouped under a common name
(saystudList). This list can be easily represented by an array
called studList for N = 45 students as shown inFigure 3.1
Fig. 3.1 Schematic representation of an array

In fact, the studList shown in Figure 3.1 can be looked upon by the
following two points of views:
1. It is a linear list of 45 names stored in contiguous locationsan abstract
view of a list having finite number of homogeneous elements (i.e., 45
names).
2. It is a set of 0 to 44 memory locations sharing a common name
called studListit is an array data structure in which 45 names have been
stored.

It may be noted that all the elements in the array are of same type,
i.e., string of char in this case. The individual elements within the array
can be designated by an index. The individual elements are randomly
accessible by integers, called the index.
For instance, the zeroth element (Sagun) in the list can be referred to as
studList [0] and the 43rd element (Ridhi) as studList [43] where 0
and 43 are the indices. An index has to be of type integer.
An index is also called a subscript. Therefore, individual elements of an
array are called subscripted variables. For instance, studList
[0] and studList[43] are subscripted variables.
From the above discussion, we can define an array as a finite
ordered collection of items of same type. It is a set of index,
value pairs.
An array is a built-in data structure in every programming language.
Arrays are designed to have a fixed size. Some languages provide zerobased indexing whereas other languages provide one-basedindexing.
C is an example of zero-based indexing language because the index of
its arrays starts from 0. Pascal is the example of one-based
indexing because the index of its arrays starts from 1.
An array whose elements are specified by a single subscript is known
as one-dimensional array. The array whose elements are specified by
two or more than two subscripts is called as multi-dimensional array.
3.2 ONE-DIMENSIONAL ARRAYS

One-dimensional arrays are suitable for processing lists of items of

identical types. They are very useful for problems that require the same
operation to be performed on a group of data items. ConsiderExample 1.
Example 1: The examination committee of the university has decided to
give 5 grace marks to every student of a class. Assume that there are 30
students in the class and their marks are stored in a list
called marksList. Write a program that adds the grace marks to
every entry in the marksList.

Solution: As the problem deals with a list of marks, we can select a onedimensional array calledmarksList to represent the list. The grace
marks are to be added to all the elements of the list, for which we must
use the For-loop. The required program is given below:
/*Thisprogramaddsgracemarkstoallelementsofa
listofmarks*/
#include<stdio.h>
voidmain()
{
intmarksList[30];
inti;
intgrace=5;/*gracemarks*/
/*ReadthemarksList*/
printf(\nEnterthelistofmarksonebyone);
for(i=0;i<30;i++)
{
scanf(%d,&marksList[i]);
marksList[i]=marksList[i]+grace;
}
/*PrinttheFinallist*/
printf(\nThefinalListis);
printf(\nS.No.\tMarks);
for(i=0;i<30;i++)
printf(\n%d\t%d,i,marksList[i]);
}
It may be noted that in the above program, a component by component
processing has been done on the all elements of the array
called marksList.
The following operations are defined on array data structure:
o Traversal: An array can be travelled by visiting its elements starting from
the zeroth element to the last in the list.
o Selection: An array allows selection of an element for a given (random)
index.
o Searching: An array can be searched for a particular value.
o Insertion: An element can be inserted in an array at a particular location.
o Deletion: An element can be deleted from a particular location of an array.
o Sorting: The elements of an array can be manipulated to obtain a sorted
array.

We shall discuss the abovementioned operations on arrays in the

subsequent sections.
3.2.1 Traversal
It is an operation in which each element of a list, stored in an array, is
visited. The travel proceeds from the zeroth element to the last element
of the list. Processing of the visited element can be done as per the
requirement of the problem.
Example 2: Write an algorithm called listTravel that travels a
list stored in an array called List of sizeN. Each visited element is
operated upon by an operator called OP.
Solution: The For-loop can be employed to visit and apply the
operator OP on every element of the array. The required algorithm is
given below:
Algorithm listTravel()
AlgorithmlistTravel()
{
Step
1.For(i=0;i<N,i++)
{
1.1List[i]=OP(List[i]);
}
}
Example 3: A list of N integer numbers is given. Write a program that
travels the list and determines as to how many of the elements of the list
are less than zero, zero, and greater than zero.
Solution: In this program, three counters called numNeg,
numZero and numPos, would be used to keep track of the number of
elements that are less than zero, zero, and greater than zero present in
the list. The required program is given below:
/*Thisprogramdeterminesthenumberoflessthan
zero,zero,andgreaterthanzeronumberspresentin
alist*/
#include<stdio.h>
voidmain()

{
intList[30];
intN;
inti,numNeg,numZero,numPos;
printf(\nEnterthesizeofthelist);
scanf(%d,&N);
printf(Entertheelementsonebyone);
/*ReadtheList*/
for(i=0;i<N;i++)
{
printf(\nEnterNumber:);
scanf(%d,&List[i]);
}
numNeg=0;/*Initializecounters*/
numZero=0;
numPos=0;
/*TraveltheList*/
for(i=0;i<N;i++)
{
if(List[i]<0)
numNeg=numNeg+1;
else
if(List[i]==0)
numZero=numZero+1;
else
numPos=numPos+1;
}
}
3.2.2 Selection
An array allows selection of an element for a given (random) index.
Therefore, the array is also called random access data structure. The
selection operation is useful in a situation wherein the user wants query
about the contents of a list for a given position.
Example 4: Write a program that stores a merit list in an array
called Merit. The index of the array denotes the merit position, i.e.,
1st, 2nd, 3rd, etc. whereas the contents of various array locations
represent the percentage of marks obtained by the candidates. On users
demand, the program displays the percentage of marks obtained for a
given position in the merit list as per the following format:

Position: 4
Percentage: 93.25
Solution: An array called Merit would be used to store the given merit
list. Two variables pos andpercentage would be used to store the merit
position and the percentage of a candidate, respectively. A do-while loop
would be used to iteratively interact with the user through following
menu displayed to the user:
Menu:
Query----1
Quit------2
Enter your choice
Note: As per the demand of the program, we will leave the zeroth
location as unused and fill the array from index 1.
The required program is given below:
/*Thisprogramillustratestherandomselection
operationallowedbyarraydatastructures*/
#include<stdio.h>
#include<conio.h>
voidmain()
{
floatMerit[30];
intsize;
inti;
intpos;
floatpercentage;
intchoice;
printf(\nEnterthesizeofthelist);
scanf(%d,&size);
printf(\nEnterthemeritlistonebyone);
for(i=1;i<=size;i++)/*ReadMerit*/
{
printf(\nEnterdata:);
scanf(%f,&Merit[i]);
}
/*displaymenuandstartsession*/
do
{
clrscr();

printf(\nMenu);
printf(\nQuery1);
printf(\nQuit2);
printf(\nEnteryourchoice:);
scanf(%d,&choice);
/*useswitchstatement
fordeterminingthechoice*/
switch(choice)
{
case1:printf(\nEnterPosition);
scanf(%d,&pos);
percentage=Merit[pos];/*therandom
selection*/
printf(\nposition=%d,pos);
printf(\npercentage=%4.2f,percentage);
break;
case2:printf(\nQuitting);
}
printf(\npressakeytocontinue);
getch();
}
while(choice!=2);
}
Searching and sorting are very important activities in a data processing
environment and they are discussed in detail in the subsequent section.
3.2.3 Searching
There are many situations where we want to find out whether a
particular item is present in a list or not. For instance, in a given voter
list of a colony a person may search his name to ascertain whether he is a
valid voter or not. For similar reasons, passengers look for their names in
the railway reservation lists.
Note: System programs extensively search symbols, literals,
mnemonics, compiler and assembler directives, etc.
In fact, search is an operation in which a given list is searched for a
particular value. The location of the searched element is
informed. Search can be precisely defined an activity of looking for a
value or item in a list.

A list can be searched sequentially wherein the search for the data item
starts from the beginning and continues till the end of the list. This
simple method of search is also called linear search. It may be noted
that for a list of size 1,000, the worst case is 1,000 comparisons.
Let us consider a situation wherein we are interested in searching a list
of numbers called numListfor an element having value equal to the
contents of a variable called val. It is desired that the location of the
element, if found, be displayed.
The list of numbers can be comfortably stored in an array
called numList of type int. To find the element, the list would be
travelled in such a manner that each visited element would be compared
to the variable val. If the match is found, then the location of the
corresponding position would be stored in a variable called Pos. As the
number of elements in the list is known, For-loop would be used to travel
the array.
Algorithm searchList()
{
step
1.readnumList
2.readVal
3.Pos=1initializePostoanonexisting
position
4.for(i=0;i<N;i++)
{
4.1if(val==numList[i])
Pos=i;
}
5.if(Pos!=1)
printPos.
}
The C implementation of the above algorithm is given below:
/*ThisprogramsearchesagivenvaluecalledValin
alistcallednumList*/
#include<stdio.h>
voidmain()
{
intnumList[20];
intN;/*Sizeofthelist*/
intPos,val,i;

printf(\nEnterthesizeoftheList);
scanf(%d,&N);
printf(\nEntertheelementsonebyone);
for(i=0;i<N;i++)
{
scanf(%d,&numList[i]);
}
printf(\nEnterthevaluetobesearched);
scanf(%d,&val);
/*Searchtheelementanditspositionin
thelist*/
Pos=1;
for(i=0;i<N;i++)
{
if(val==numList[i])
{
Pos=i;
break;
}
/*Theelementisfoundcomeoutof
loop*/
}
if(Pos!=1)
printf(\nTheelementfoundat%dlocation,
Pos);
else
printf(\nSearchFailed);
}
Example 5: Write a program that finds the second largest element in a
given list of N numbers.
Solution: The most basic solution for this problem is that if there are
two elements in the list then the smaller of the two will be the second
largest. In this case, we would set the second largest number to 9999,
a value not possible in the list.
In given problem, the list will be searched linearly for the required
second largest number. Two variables
called firstLarge and secLarge would be employed to store the first
and second largest numbers. The following algorithm will be used:
Algorithm SecLarge()

{
step
1.readnum;
2.firstLarge=num;
3.secLarge=9999;
4.for(i=2;i<5N;i++)
{
4.1readnum;
4.2if(firstLarge<num)
{secLarge=firstLarge;
firstLarge=num;
}
else
if(secLarge<num)
secLarge=num;
}
4.promptSecondLarge5;
5.writesecLarge;
}
The equivalent program is given below:
/*Thisprogramfindsthesecondlargestinagiven
listofnumbers*/
#include<stdio.h>
intNum,firstLarge,secLarge;
intN,i;
voidmain()
{
printf(\nEnterthesizeofthelist);
scanf(%d,&N);
printf(\nEnterthelistonebyone);
scanf(%d,&Num);
firstLarge=Num;
secLarge=9999;
for(i=2;i<5N;i++)
{
scanf(%d,&Num);
if(firstLarge<Num)
{secLarge5firstLarge;
firstLarge5Num;

}
else
if(secLarge<Num)
secLarge=Num;
}
printf(\nSecondLarge=%d,secLarge);
}
The above discussed search through a list, stored in an array, has the
following characteristics:
o The search is linear.
o The search starts from the first element and continues in a sequential
fashion from element to element till the desired entry is found.
o In the worst case, a total number of N steps need to be taken for a list of size
N.

Thus, the linear search is slow and to some extent inefficient. In special
circumstances, faster searches can be applied.
For instance, binary search is a faster method as compared to linear
search. It mimics the process of searching a name in a directory wherein
one opens a page in the middle of the directory and examines the page
for the required name. If it is found, the search stops; otherwise, the
search is applied either to first half of the directory or to the second half.
3.2.3.1 Binary Search
If a list is already sorted, then the search for an entry (sayVal) in the
list can be made faster by using divide and conquer technique. The list
is divided into two halves separated by the middle element as shown
in Figure 3.2.

Fig. 3.2 Binary search

The binary search follows the following steps:
Step
1. The middle element is tested for the required entry. If found, then its
position is reported else the following test is made.
2. If Val < middle, search the left half of the list, else search the right half of
the list.

3. Repeat step 1 and 2 on the selected half until the entry is found otherwise
report failure.

This search is called binary because in each iteration, the given list is
divided into two (i.e., binary) parts. Therefore, in next iteration the
search becomes limited to half the size of the list to be searched. For
instance, in first iteration the size of the list is N which reduces to almost
N/2 in the second iteration and so on.
Let us consider a sorted list stored in an array called Series given
in Figure 3.3.
Fig. 3.3 The Series containing nine elements
Suppose we desire to search the Series for a value (say14) and its
position in it. Binary search begins by looking at the middle value in the
Series. The middle index of the array is approximated by averaging the
first and last indices and truncating the result, i.e., (0 + 8)/2 = 4. Now,
the content of the fourth location in Series happens to be 11 as shown
in Figure 3.4. Since the value we are looking for (i.e., 14) is greater than
11, the middle value, it may be present in the right half (Series [5]
to Series[8]).

Fig. 3.4 The middle value in the Series (1st step)

Now the middle of the right half is approximated i.e., (5 + 8)/2 = 6. We
find that the desired element exists at the middle of the right half,
i.e., Series [6] = 14 as shown in Figure 3.5.

Fig. 3.5 The middle value in the Series (2nd step)

It may be noted that the desired element has been found only in two
steps. Thus, it is a much faster method as compared to the linear search.
For instance, a list of 1,000 sorted elements would require 10
comparisons to search the entire list. An algorithm for this method is
given below.
In this algorithm, we would employ a Boolean variable called flag.
The flag will indicate the presence or absence of the element being
searched.

Algorithm binSearch()
{
Step
1.First=0;
2.Last=N1;
3.Pos=1;
4.Flag=false;
5.While(First<=LastandFlag==false)
{
5.1Middle=(first+last)div2
if(Series[middle]==Val)
{Pos=middle;
flag=true;
breakfromtheloop;
}
else
if(Series[middle]<Val)First=middle
+1;
elseLast=middle1;
}
6.if(flag==true)
promptThevaluefoundat;
writepos;
elsepromptThevaluenotfound;
}
It may be noted that div operator has been used to indicate that it is
an integer division. The integer division will truncate the results to the
nearest integer. If the desired value is found, then the flag is set to true
and the while loop terminates; otherwise, a stage arrives when first
becomes greater than the last, indicating the failure of the search. Thus,
the variables First and Last keep track of the lower and the upper
bounds of the array, respectively.
Example 6: Write a program that uses binary search to search a given
value called Val in a list of Nnumbers called Series.
Solution: The Algorithm binSearch() discussed above is used to
write the required program.
/*Thisprogramusesbinarysearchtofindagiven
valuecalledvalinalistofNnumbers*/

#include<stdio.h>
#definetrue1
#definefalse0
voidmain()
{
intFirst;
intLast;
intMiddle;
intSeries[20];/*ThelistofNsortednumbers*/
intVal;
intflag;/*Thevaluetobesearched*/
intN,Pos,i;
printf(\nEnterthesizeofthelist);
scanf(%d,&N);
printf(\nEnterthesortedlistonebyone);
for(i=0;i<N;i++)
{
scanf(%d,&Series[i]);
}
printf(\nEnterthenumbertobesearched);
scanf(%d,&Val);
/*BINSEARCHbegins*/
Pos=1;/*Nonexistingposition*/
flag=false;/*Assumesearchfailure*/
First=0;
Last=N1;
while((First<=Last)&&(flag==false))
{
Middle=(First+Last)/2;
if(Series[Middle]==Val)
{Pos=Middle;
flag=true;
break;
}
else
if(Series[Middle]<Val)
First=Middle+1;
else
Last=Middle1;
}
if(flag==true)

printf(\nThevaluefoundat%d,Pos);
else
printf(\nThevaluenotfound);
}
Binary search through a list, stored in an array, has the following
characteristics:
o The list must be sorted, i.e., ordered.
o It is faster as compared to the linear search.
o A list with large number of elements would increase the total execution
time, the reason being that the list must be ordered which requires extra
effort.

3.2.3.2 Analysis of Binary Search

As discussed above, we know that the binary search took two steps to
search an element in a list of nine elements. However, in the worst case
scenario for a list of 32 elements, we can visualize as follows:
1st step would get a sublist of size 16
2nd step would get a sublist of size 8
3rd step would get a sublist of size 4
4th step would get a sublist of size 2
5th step would get a sublist of size 1
Let us tabulate (Table 3.1) the number of steps taken by binary search for
a given list of size n. Fromtable 3.1, we can infer that for a given input
size n = 2k, number of steps = k.
Table 3.1
Size of list (n)

Number of steps

(23)

32

(25)

256

(28)

512

(29)

Now the relation n = 2k can be rewritten as k = log2n

(ax = n x = loga n)

Thus, the time taken by binary search T(n) = k = log2n

To satisfy the relation 2.1, the above relation can be rewritten as shown
below:
T(n) <= 1*log2n

where c = 1 and n0 = 0.

Comparing the above relation with relation 2.1, we get g(n) = log2n.
Therefore, the above relation can be expressed as:
T(n) = O(log2n)
Hence, binary search has time complexity of the order of log2n, i.e.,
O(log2n).
3.2.4 Insertion and Deletion
3.2.4.1 Insertion
Insertion is an operation in which a new element is added at a particular
place in a list. When the list is represented using an array, the element
can be added very easily at the end of the list provided the space is
available. However, insertion of an element at a particular location
within the list is a difficult operation. For instance, if it is desired to
insert an element at ith location of a list then an empty space will have to
be created at the ith location of the list. In fact, all the elements from ith
location are shifted by one place towards right to accommodate the new
element.

Fig. 3.6 The list charList before insertion

Consider a list of characters, represented using an array
called charList as shown in Figure 3.6.Suppose it is desired to
insert a character D at its proper place in charList, i.e., at the 3rd
location.
Now to accommodate the character D, the characters G, F, and E will
have to be shifted one step towards right to locations 6th, 5th, and 4th,
respectively as per the operations shown below:
charList[6] = charList[5];
charList[5] = charList[4];
charList[4] = charList[3];

Once the above copying operation is over, charList [3] becomes

available for the storage of incoming character D as shown below:
charList[3]=D;
Fig. 3.7 The array charList after insertion
The array charList after the insertion operation is shown in Figure 3.7.
An algorithm for insertion of an element called val at ith location in an
array called charList of size N is given below. Let us assume that the
last element in the array is at Mth location. For instance, N = 8 and M =
5 in case of charList shown in Figure 3.6. A variable called Back is
being used to point to the last empty location:
Algorithm insertArray()
{
step
/*pointtothelastvacantposition*/
1.if(M<N)thenBack=M+1;
else
STOP;
2.while(Back>I);
{/*Copyelementstonextlocationinthelist*/
2.1charList[Back]=charList[Back1];
2.2Back=Back1;
}
3.charList[I]=val;/*Inserttheelementatith
location*/
4.M=M+1;
}
Example 7: Write a program that inserts an element called key at a
location called loc in a list of M numbers called List.
Solution: The algorithm insertArray() is being used to implement
the required program.
/*ThisprograminsertsanelementcalledKeyintoa
listofnumbers*/
#include<stdio.h>
#defineN20

voidmain()
{
intList[N];
intKey;
intLoc;
inti,M;
intback;
printf(\nEnterthesizeofthelist(<20));
scanf(%d,&M);
printf(\nEnterthelistonebyone);
for(i50;i<M;i++)
{
scanf(%d,&List[i]);
}
printf(\nEnterthekeyandlocationrelativeto0
index);
scanf(%d%d,&Key,&Loc);
/*InsertthekeyatLocinList*/
if(Loc>M+1)
printf(\nInsertionnotpossible);
else
{
back=M+1;
/*Shiftelementsonestepright*/
while(back>Loc)
{
List[back]=List[back1];
back;
}
/*InsertKey*/
List[back]=Key;
M=M+1;
/*Displaythefinallist*/
printf(\nTheFinalListis);
for(i=0;i<M;i++)
{
printf(%d,List[i]);
}
}
}

3.2.4.2 Deletion
Deletion is the operation that removes an element from a given location
of a list. When the list is represented using an array, the element can be
very easily deleted from the end of the list. However, if it is desired to
delete an element from an ith location of the list, then all elements from
the right of (i+1)th location will have to be shifted one step towards left
to preserve contiguous locations in the array.
For instance, if it is desired to remove an element from 4th location of
the list given in Figure 3.8, then all elements from right of 4th location
would have to be shifted one step towards left.

Fig. 3.8 Delete element from 4th location

Now to delete the character E, the characters F and G will have to
be shifted one step towards left to locations 4th and 5th, respectively as
per the operations shown below:
charList[4]=charList[5];
charList[5]=charList[6];
It may be noted that the contents of charList[5] (i.e., F)
overwrites the contents of charList[4] (i.e., E). The
array charList after the deletion operation is shown in Figure 3.9.
Fig. 3.9 The array charList after deletion
An algorithm for deletion of an element from ith location in an array
called charList of size N is given below. Let us assume that the last
element in the array is at Mth location. For instance, N = 8 and M = 6 in
case of charList shown in Figure 3.8. A variable called Back is used to
point to the last empty location.
Algorithm delElement()
{
Step
1.Back=I;/*pointtothelocationfromwhere
deletionisdesired*/
2.while(Back<M)

{
2.1charList[Back]=charList[Back+1];/*Shift
elementsone
stepleft*/
2.2Back=Back+1;
}
3.M=M1;
4.Stop
}
It may be noted that in step 3, the contents of M have been decremented
by 1 to indicate that after deletion the last element in
the charList would be at (M - 1)th location.
Example 8: Write a program that deletes an element from a location
called loc in a list of M numbers called List.
Solution: The algorithm delElement() is used to implement the
required program.
/*Thisprogramdeletesanelementfromlistof
numbers*/
#include<stdio.h>
#defineN20
voidmain()
{
intList[N];
intLoc;
inti,M;
intback;
printf(\nEnterthesizeofthelist(<20));
scanf(%d,&M);
printf(\nEnterthelistonebyone);
for(i=0;i<M;i++)
{
scanf(%d,&List[i]);
}
printf(\nEnterlocationfromwherethedeletionis
required);
scanf(%d,&Loc);
/*InsertthekeyatLocinList*/
if(Loc>M)

printf(\nDeletionnotpossible);
else
{
back=Loc;
/*Shiftelementsonestepleft*/
while(back<M)
{
List[back]=List[back+1];
back++;
}
M=M1;
/*Displaythefinallist*/
printf(\nTheFinalListis.);
for(i=0;i<M;i++)
{
printf(%d,List[i]);
}
}
}
It may be noted that arrays are not suitable data structures for problems
requiring insertions and deletions in a list. The reason is that we need to
shift elements to right or to left for insertion and deletion operations,
respectively. The problem aggravates when the size of the list is very
large as an equally large number of shift operations would be required to
insert and delete the elements. Thus, insertion and deletion are very slow
operations as far as arrays are concerned.
3.2.5 Sorting
It is an operation in which all the elements of a list are arranged in a
predetermined order. The elements can be arranged in a sequence from
smallest to largest such that every element is less than or equal to its
next neighbour in the list. Such an arrangement is called ascending
order. Assuming an array called List containing N elements, the
ascending order can be defined by the following relation:
List[i]<=List[i+1],0<i<N1
Similarly in descending order, the elements are arranged in a sequence
from largest to smallest such that every element is greater than or equal
to its next neighbour in the list. The descending order can be defined by
the following relation:

List[i]>=List[i+1],0<i<N1
It has been estimated that in a data processing environment, 25 per cent
of the time is consumed in sorting of data. Many sorting algorithms have
been developed. Some of the most popular sorting algorithms that can be
applied to arrays are in-place sort algorithms. An in-place algorithm is
generally a comparison-based algorithm that stores the sorted elements
of the list in the same array as occupied by the original one. A detailed
discussion on sorting algorithms is given in subsequent sections.
3.2.5.1 Selection Sort
It is a very simple and natural way of sorting a list. It finds the smallest
element in the list and exchanges it with the element present at the head
of the list as shown in Figure 3.10.

Fig. 3.10 Selection sort (first pass)

It may be noted from Figure 3.10 that initially, whole of the list was
unsorted. After the exchange of smallest with the element on the head of
the list, the list is divided into two parts: sorted and unsorted.
Now the smallest is searched in the unsorted part of the list, i.e., 2 and
exchanged with the element at the head of unsorted part, i.e., 20 as
shown in Figure 3.11.

Fig. 3.11 Selection sort (second pass)

This process of selection and exchange (i.e., a pass) continues in this
fashion until all the elements in the list are sorted (see Figure 3.12).
Thus, in selection sort, two steps are importantselection andexchange.

From Figures 3.11 and 3.12, it may be observed that it is a case of nested
loops. The outer loop is required for passes over the list and the inner
loop for searching smallest element within the unsorted part of the list.
In fact, for N number of elements, N 1 passes are made.
An algorithm for selection sort is given below. In this algorithm, the
elements of a list stored in an array called LIST[N] are sorted in
ascending order. Two variables called Small and Pos are used to locate
the smallest element in the unsorted part of the list. Temp is the variable
used to interchange the selected element with the first element of the
unsorted part of the list.
Algorithm SelSort()
{
Step
1.ForI=1toN1/*OuterLoop*/
{
1.1small=List[I];
1.2Pos=I;
1.3ForJ=I+1toN/*InnerLoop*/
{
1.3.1if(List[J]<small)
{
small=List[J];
Pos=J;/*Notethepositionofthesmallest*/
}
}
1.4Temp=List[I];/*Exchangesmallestwiththe
Head*/
1.5List[I]=List[Pos];
1.6List[Pos]=Temp;
}
2.Printthesortedlist
}

Fig. 3.12 Selection sort

Example 9: Given is a list of N randomly ordered numbers. Write
program that sorts the list in ascending order by using selection sort.
Solution: The required program is given below:
In this program, the elements of a list are stored in an array
called List. The elements are sorted using above given
Algorithm selSort(). Two variables small and pos have been used to
locate the smallest element in the unsorted part of the list. Temp is a
variable used to interchange the selected element with the first element
of the unsorted part of the list. With each step, the unsorted part
becomes smaller. The process is repeated till all the elements are sorted.

/*Thisprogramsortsalistbyusingselectionsort
*/
#include<stdio.h>
main()
{
intlist[10];
intsmall,pos,N,i,j,temp;
printf(\nEnterthesizeofthelist:);
scanf(%d,&N);
printf(\nEnterthelist:);
for(i50;i<N;i++)
{
printf(\nEnterNumber:);
scanf(%d,&list[i]);
}
/*Sortthelist*/
for(i50;i<N1;i++)
{
small=list[i];
pos=i;
/*Findthesmallestoftheunsortedlist*/
for(j=i+1;j<N;j++)
{
if(small>list[j])
{
small=list[j];
pos=j;
}
}
/*Exchangethesmallwiththe
firstelementofunsortedlist*/
temp=list[i];
list[i]=list[pos];
list[pos]=temp;
}
printf(\nThesortedlist);
for(i50;i<N;i++)
printf(%d,list[i]);
}

3.2.5.2 Analysis of Selection Sort

In selection sort, there are two major operations: comparison and
exchange. The average number of exchange operations would be difficult
to estimate. Therefore, we will focus on comparison operations.
It may be noted that for every execution of outer loop, the inner loop
executes in decreasing order. For example, in a list of (say) eight
numbers, the first number would be compared to the remaining seven
numbers to find out the smallest. After bringing the smallest to the first
position, the second number would be compared to the remaining six,
the third number with remaining five, and so on. Thus, the total number
of comparisons for a list on N elements would be:
(N 1) + (N 2) + .2 + 1 = (N 1) N/2 = N2/2 N/2
Thus, T(n) = N2/2 N/2.
we can say that
N2/2 N/2 <= N2/2.
Therefore, we can say that
T(n) <= N2/2

where c = , n0= 0 , g(n) = N2

T(n) = O(N2)
Hence, selection sort has time complexity of the order of N2, i.e., O(N2).
3.2.5.3 Bubble Sort
It is also a very simple sorting algorithm. It proceeds by looking at the
list from left to right. Each adjacent pair of elements is compared.
Whenever a pair is found not to be in order, the elements are exchanged.
Therefore after the first pass, the largest element bubbles up to the right
end of the list. A trace of first pass on a list of numbers is shown
in Figure 3.13.
It may be noted that after the pass is over, the largest element in the list
(i.e., 20) has bubbled up to the end of the list and six exchanges were
made. Now the same process can be repeated for the list for second pass
as shown in Figure 3.14.
We observe that after the second pass is over, the list has become sorted
and only two exchanges were made. Now a point worth noting is as to
how and when it will be decided that the list has become sorted. The
simple criteria would be to check whether or not any exchange(s) has
been made during the current pass. If yes, then the list is not yet sorted

otherwise if it is no, then it can be decided that the list has just become
sorted.
An algorithm for bubble sort is given below. In this algorithm, the
elements of a list, stored in an array called List[N], are sorted in an
ascending order. The algorithm uses two loopsthe outer while loop and
inner For loop. The inner For loop makes a pass on the list. If during the
pass, any exchange(s) is made then it is recorded in a variable
called flag, i.e., flag is set to false. The outer while loop keeps track of
the flag. As soon as the flag informs that no exchange(s) took place
during the current pass indicating that the list is now sorted, the
algorithm stops.

Fig. 3.13 First pass of bubble sort on a list of numbers

Fig. 3.14 Second pass of the bubble sort

Algorithm bubbleSort()
{
Step
1.Flag=false;
2.While(Flag==false)
{
2.1Flag=true;
2.2Forj=0toN2
{
2.2.1if(List[J]>List[J+1])
{
temp=List[J];
List[J]=List[J+1];
List[J+1]=temp;
Flag=false;
}
}
}
3.Printthesortedlist
4.Stop
}
It may be noted that the algorithm stops as soon as the list becomes
sorted. Thus, bubble sort is a very useful algorithm when the list is
almost sorted i.e., only a very small percentage of elements are out of
order.
Example 10: Given is a list of N randomly ordered numbers. Write a
program that sorts the list in ascending order by using bubble sort.
Solution: The required program is given below:
In this program, the elements of a list are stored in an array
called List. The elements are sorted using above given
algorithm bubbleSort().
/*Thisprogramsortsagivenlistofnumbersin
ascendingorder,usingbubblesort*/
#include<stdio.h>
#defineN20
#definetrue1
#definefalse0
voidmain()

{
intList[N];
intflag;
intsize;
inti,j,temp;
intcount;/*countsthenumberofpasses*/
printf(\nEnterthesizeofthelist(<20));
scanf(%d,&size);
printf(\nEnterthelistonebyone);
for(i=0;i<size;i++)
{
scanf(%d,&List[i]);
}
/*Sortthelistbybubblesort*/
flag=false;
count=0;
while(flag==false)
{
flag=true;/*Assumenoexchangetakesplace*/
count++;
for(j=0;j<size1;j++)
{
if(List[j]>List[j+1])
{/*Exchangethecontents*/
temp=List[j];
List[j]=List[j+1];
List[j+1]=temp;
flag=false;/*Recordtheexchange
operation*/
}
}
}
/*Printthesortedlist*/
printf(\nThesortedlistis.);
for(i=0;i<size;i++)
printf(%d,List[i]);
printf(\nThenumberofpassesmade=%d,
count);
}

It may be noted that the above program has used a variable called count
that counts the number of passes made while sorting the list. The test
runs conducted on the program have established that the program is
very efficient in case of almost sorted list of elements. In fact, it takes
only one scan to establish that the supplied list is already sorted.
Note: Bubble sort is also called a sinking sort meaning that the
elements sink down in the list to their proper position.
3.2.5.4 Analysis of Bubble Sort
In bubble sort, again, there are two major operationscomparison and
exchange. The average number of exchange operations would be difficult
to estimate. Therefore, we will focus on comparison operations.
A closer look reveals that for a list of N numbers, bubble sort also has the
following number of comparisons, i.e., same as the selection sort.
(N 1) + (N 2) + . 2 + 1 = (N 1) N/2 = N2/2 N/2
Therefore, the time complexity of bubble sort is also O(N2).
However if the list is already sorted in the ascending order, no exchange
operations would be required and it becomes the best case. The
algorithm will have only N comparisons, i.e., only a linear running time.
3.2.5.5 Insertion Sort
This algorithm mimics the process of arranging a pack of playing cards.
In the pack of cards, the first two cards are put in correct relative order.
The third card is inserted at correct place relative to the first two. The
fourth card is inserted at the correct place relative to the first three, and
so on.
Given a list of numbers, it divides the list into two partsorted part and
unsorted part. The first element becomes the sorted part and the rest of
the list becomes the unsorted part as shown in Figure 3.15. It picks up
one element from the front of the unsorted part and inserts it at its
proper position in the sorted part of the list. This insertion action is
repeated till the unsorted part is exhausted.

Fig. 3.15 Insertion sort

It may be noted that the insertion operation requires following steps:
Step
1. Scan the sorted part to find the place where the element, from unsorted
part, can be inserted. While scanning, shift the elements towards right to
create space.
2. Insert the element, from unsorted part, into the created space.

The algorithm for the insertion sort is given below. In this algorithm, the
elements of a list stored in an array called List[N] are sorted in an
ascending order. The algorithm uses two loopsthe outer For loop and
inner while loop. The inner while loop shifts the elements of the sorted
part by one step to right so that proper place for incoming element is
created. The outer For loop inserts the element from unsorted part into
the created place and moves to next element of the unsorted part.
Algorithm insertSort()
{
Step
1.ForI=2toN/*Thefirstelementbecomesthe
sortedpart*/
{
1.1Temp=List[I];/*Savetheelementfrom
unsortedpartintotemp*/
1.2J=I1;
1.3While(Temp<=List[J]ANDJ>=0)
{
List[J+1]=List[J];/*Shiftelements
towardsright*/
J=J1;

}
1.4List[J+1]=Temp;
}
2.Printthelist
3.Stop
}
Example 11: Given is a list of N randomly ordered numbers. Write a
program that sorts the list in ascending order by using insertion sort.
Solution: The required program is given below:
/*Thisprogramsortsagivenlistofnumbersin
ascendingorderusinginsertionsort*/
#include<stdio.h>
#defineN20
voidmain()
{
intList[N];
intsize;
inti,j,temp;
printf(\nEnterthesizeofthelist(<20));
scanf(%d,&size);
printf(\nEnterthelistonebyone);
for(i=0;i<size;i++)
{
scanf(%d,&List[i]);
}
/*SortthelistbyInsertionsort*/
for(i=1;i<size;i++)
{
temp=List[i];/*Pickandsavethefirstelement
oftheunsortedpart*/
j=i1;
while((temp<List[j])&&(j>=0))/*Scanfor
properplace*/
{
List[j+1]=List[j];
j=j1;
}
List[j+1]=temp;/*Inserttheelementatthe
properplace*/

}
/*Printthesortedlist*/
printf(\nThesortedlistis.);
for(i=0;i<size;i++)
{
printf(%d,List[i]);
}
}
3.2.5.6 Analysis of Insertion Sort
A critical look at the algorithm indicates that in worst case, i.e., when the
list is sorted in reverse order, the jth iteration requires (j 1)
comparisons and copy operations. Therefore, the total number of
comparison and copy operations for a list of N numbers would be:
1 + 2 + 3 +.(N 2) + (N 1) = (N 1) N/2 = N2/2 N/2
Therefore, the time complexity of insertion sort is also O(N2).
However if the list is already sorted in the ascending order, no copy
operations would be required. It becomes the best case and the
algorithm will have only N comparisons, i.e., a linear running time.
3.2.5.7 Merge Sort
This method uses following two concepts:
1. If a list is empty or it contains only one element, then the list is already
sorted. A list that contains only one element is also called singleton.
2. It uses the old proven technique of divide and conquer to recursively
divide the list into sub-lists until it is left with either empty or singleton
lists.

In fact, this algorithm divides a given list into two almost equal sub-lists.
Each sub-list, thus obtained, is recursively divided into further two sublists and so on till singletons or empty lists are left as shown in Figure
3.16.

Fig. 3.16 First step of merge sort (divide)

Since the singletons and empty lists are inherently sorted, the only step
left is to merge the singletons into sub-lists containing two elements each
(see Figure 3.17)which are further merged into sub-lists containing four
elements each and so on. This merging operation is recursively carried
out till a final merged list is obtained as shown in Figure 3.17.

Fig. 3.17 Second step of merge sort (merge)

Note: The merge operation is a time consuming and slow operation. The
working of merge operation is discussed in the next section.
Merging of lists It is an operation in which two ordered lists are
merged into a single ordered list. The merging of two lists PAR1
and PAR2 can be done by examining the elements at the head of the two
lists and selecting the smaller of the two. The smaller element is then
stored into a third list calledmergeList. For example, consider the
lists PAR1 and PAR2 given in Figure 3.18. Let Ptr1, Ptr2, and Ptr
3 variables point to the first locations of lists PAR1, PAR2, and PAR3,
respectively. The comparison of PAR1[Ptr1] and PART2 [Ptr2]
shows that the element of PAR1 (i.e., 2) is smaller. Thus, this element
will be placed in the mergeList as per the following operation:
mergeList[Ptr3]=PAR1[Ptr1];
Ptr1++;
Ptr3++;

Fig. 3.18 Merging of lists (first step)

Since an element from the list PAR1 has been taken to mergeList, the
variable Ptr1 is accordingly incremented to point to the next location in

the list. The variable Ptr3 is also incremented to point to next vacant
location in mergeList.
This process of comparing, storing and shifting is repeated till both the
lists are merged and stored inmergeList as shown in Figure 3.19.

Fig. 3.19 Merging of lists (second step)

It may be noted here that during this merging process, a situation may
arise when we run out of elements in one of the lists. We must, therefore,
stop the merging process and copy rest of the elements from unfinished
list into the final list.
The algorithm for merging of lists is given below. In this algorithm, the
two sub-lists are part of the same array List[N]. The first sub-list is
stored in locations List[lb] to List[mid] and the second sub-list is
stored in locations List[mid+1] to List[ub] where lband
ub mean lower and upper bounds of the array, respectively.
Algorithm merge(List,lb,mid,ub)
{
Step
1.ptr1=lb;/*indexoffirstlist*/
2.ptr2=mid;/*indexofsecondlist*/
3.ptr3=lb;/*indexofmergedlist*/
4.while((ptr1<mid)&&ptr2<=ub)/*mergethe
lists*/
{
4.1if(List[ptr1]<=List[ptr2])
{mergeList[ptr3]=List[ptr1];/*elementfrom
firstlististaken*/
ptr1++;/*movetonextelementinthelist*/
ptr3++;
}
4.2else
{mergeList[ptr3]=List[ptr2];/*elementfrom
secondlististaken*/

ptr2++;/*movetonextelementinthelist*/
ptr3++;
}
}
5.while(ptr1<mid)/*copyremainingfirstlist
*/
{
5.1mergeList[ptr3]=List[ptr1];
5.2ptr1++;
5.3ptr3++;
}
6.while(ptr2<=ub)/*copyremainingsecondlist
*/
{
6.1mergeList[ptr3]=List[ptr2];
6.2ptr2++;
6.3ptr3++;
}
7.for(i=lb;i<ptr3;i++)/*copymergedlist
backintooriginallist*/
7.1List[i]5mergeList[i];
8.Stop
}
It may be noted that an extra temporary array called mergeList is
required to store the intermediate merged sub-lists. The contents of
the mergeList are finally copied back into the original list.
The algorithm for the merge sort is given below. In this algorithm, the
elements of a list stored in an array called List[N] are sorted in an
ascending order. The algorithm has two partsmergeSort and merge.
The merge algorithm, given above, merges two given sorted lists into a
third list, which is also sorted. The mergeSort algorithm takes a list and
stores into an array called List[N]. It uses two variables lb and ub to
keep track of lower and upper bounds of list or sub-lists as the case may
be. It recursively divides the list into almost equal parts till singletons or
empty lists are left. The sublistsare recursively merged through
merge algorithm to produce final sorted list.
Algorithm mergeSort(List,lb,ub)
{
Step
1.if(lb<ub)

{
1.1mid=(lb+ub)/2;/*dividethelistinto
twosublists*/
1.2mergeSort(List,lb,mid);/*sorttheleft
sublist*/
1.3mergeSort(List,mid+l,ub);/*sortthe
rightsublist*/
1.4merge(List,lb,mid+l,ub);/*mergethelists
*/
}
2.Stop
}
Example 12: Given is a list of N randomly ordered numbers. Write a
program that sorts the list in ascending order by using merge sort.
Solution: The required program uses both the algorithms
mergeSort() and merge().
/*Thisprogramsortsagivenlistofnumbersin
ascendingorderusingmergesort*/
#include<stdio.h>
#include<conio.h>
#defineN20
voidmergeSort(intList[],intlb,intub);
voidmerge(intList[],intlb,intmid,intub);
voidmain()
{
intList[N];
inti,size;
intmid;
printf(\nEnterthesizeofthelist(<20));
scanf(%d,&size);
printf(\nEnterthelistonebyone);
for(i50;i<size;i++)
{
scanf(%d,&List[i]);
}
/*Sortthelistbymergesort*/
mergeSort(List,0,size1);
printf(\nThesortedlistis.);
for(i=0;i<size;i++)

{
printf(%d,List[i]);
}
}
voidmergeSort(intList[],intlb,intub)
{
intmid;
if(lb<ub)
{
mid=(lb+ub)/2;
mergeSort(List,lb,mid);
mergeSort(List,mid+1,ub);
merge(List,lb,mid+1,ub);
}
}
voidmerge(intList[],intlb,intmid,intub)
{
intmergeList[20];
intptr1,ptr2,ptr3;
inti;
ptr1=lb;
ptr2=mid;
ptr3=lb;
while((ptr1<mid)&&ptr2<=ub)
{
if(List[ptr1]<=List[ptr2])
{mergeList[ptr3]=List[ptr1];
ptr1++;
ptr3++;
}
else
{mergeList[ptr3]=List[ptr2];
ptr2++;
ptr3++;
}
}
while(ptr1<mid)
{mergeList[ptr3]=List[ptr1];
ptr1++;
ptr3++;
}

while(ptr2<=ub)
{mergeList[ptr3]=List[ptr2];
ptr2++;
ptr3++;
}
for(i=lb;i<ptr3;i++)
{
List[i]=mergeList[i];
}
}
3.2.5.8 Analysis of Merge Sort
The merge sort requires the following operations:
1.
2.
3.
4.

Divide the list into two sub-lists.

Sort each sub-list.
Merge the two sub-lists.
If number of elements = 1, then list is sorted.

For a list of N elements, steps 1 and 2 require two times the number of
comparisons needed for sorting a sub-list of size N/2. The step 3 requires
N steps to merge the two sub-lists.
Therefore, number of comparisons in merge sort can be defined by the
following recursive relation:
NumComp (N) = 0 if N = 1,
= 2*NumComp(N/2) + N for N > 1
Thus, NumComp (1) = 0
NumComp (N) = 2*NumComp(N/2) + N

(3.1)

By similarity, we can define the following:

NumComp (N/2) = 2*NumComp(N/4) + N
Putting above expressions into Eq. 3.1, we get
NumComp (N) = 4 NumComp(N/4) + 2N = 22NumComp(N/22) +
2N
(3.2)
By repeating the above pattern, we shall arrive at the following relation:
NumComp (N) = 8 NumComp(N/8) + 3N = 23NumComp(N/23) +
3N
(3.3)

We can generalize the above relation as:

NumComp (N) = 2kNumComp(N/2k) + kN

(3.4)

The merge sort stops when we are left with one element per partition,
i.e., when N= 2k.
Now for N = 2k or k = log N, we can rewrite Eq. 3.4 as:
NumComp
(N)

NNumComp(N/N) + N*log2N

NNumComp(1) + N*log2N = N*0 + log2N*N =

N*log2N

N*log2N

Thus, T(N) = N*log2N <= C*N*log2N + n0

T(N) = O(N*log2N) for c = 1 and n0 = 0
Thus, the time complexity of merge sort is of the order of N*log 2N or
simply n log n.
3.2.5.9 Quick Sort
This method also uses the technique of divide and conquer. On the
basis of a selected element (pivot) from of the list, it partitions the rest of
the list into two partsa sublist that contains elements less than the
pivot and other sublist containing elements greater than the pivot.
The pivot is inserted between the two sublists. The algorithm is
recursively applied to the sublists until the size of each sub
list becomes 1, indicating that the whole list has become sorted.
Consider the list given in Figure 3.20. Let the first element (i.e., 8) be the
pivot. Now the rest of the list can be divided into two partsa sub
list that contains elements less than 8 and the other sublistthat
contains elements greater than 8 as shown in Figure 3.20.

Fig. 3.20 Quick sort

Now this process can be recursively applied on the two sublists to
completely sort the whole list. For instance, 7 becomes the pivot for
left sublist and 19 becomes pivot for the right sublist.
Note: Two sublists can be safely joined when every element in the
first sublist is smaller than every element in the second sub
list. Since join is a faster operation as compared to a merge
operation, this sort is rightly named as a quick sort.
The algorithm for the quick sort is given below:
In this algorithm, the elements of a list, stored in an array
called List[N], are sorted in an ascending order. The algorithm has two
partsquickSortand partition. The partition algorithm divides the
list into two sublists around a pivot. The quickSort algorithm
takes a list and stores it into an array called List[N]. It uses two
variables lb and ub to keep track of lower and upper bounds of list
orsublists as the case may be. It employs partition algorithm to sort
the sublists.
Algorithm quickSort()
{
Step
1.Lb=0;/*setlowerbound*/
2.ub=N1;/*setupperbound*/
3.pivot=List[lb];

4.lb++;
5.partition(pivot,List,lb,ub);
}
Algorithm partition(pivot,List,lb,ub)
{
Step
1.i=lb;
2.j=ub;
3.while(i<=j)
{
/*travelthelistfromlbtillanelement
greaterthanthepivotisfound*/
3.1while(List[i]<=pivot)i++;
/*travelthelistfromubtillanelement
smallerthanthepivotisfound*/
3.2while(List[j]>pivot)j;
3.3if(i<=j)/*exchangetheelements*/
{
temp=List[i];
List[i]=List[j];
List[j]=temp;
}
}
4.temp=List[j];/*placethepivotatmidofthe
sublists*/
5.List[j]=List[lb1];
6.List[lb1]=temp;
7.if(j>lb)quicksort(List,lb,j1);/*sort
leftsublist*/
8.if(j<ub)quicksort(List,j+1,ub);/*sort
therightsublist*/
}
Example 13: Given is a list of N randomly ordered numbers. Write a
program that sorts the list in ascending order by using quick sort.
Solution: The required program uses both the algorithms
quickSort() and partition(). In this program, a variable
called Key has been used that acts as a pivot.

/*Thisprogramsortsagivenlistofnumbersin
ascendingorder,usingquicksort*/
#include<stdio.h>
#include<conio.h>
#defineN20
voidpartition(intKey,intList[],intlb,intub);
voidquicksort(intList[],intlb,intub);
voidmain()
{
intList[N];
inti,size,Pos,temp;
intlb,ub;
printf(\nEnterthesizeofthelist(<20));
scanf(%d,&size);
printf(\nEnterthelistonebyone);
for(i=0;i<size;i++)
{
scanf(%d,&List[i]);
}
/*Sortthelistbyquicksort*/
quicksort(List,0,size1);
/*Printthesortedlist*/
printf(\nThesortedlistis.);
for(i=0;i<size;i++)
{
printf(%d,List[i]);
}
}
voidquicksort(intList[],intlb,intub)
{
intKey;/*ThePivot*/
Key=List[lb];lb++;
partition(Key,List,lb,ub);
}
voidpartition(intKey,intList[],intlb,intub)
{
inti,j,temp;
i=lb;
j=ub;
while(i<=j)
{

while(List[i]<=Key)i++;
while(List[j]>Key)j;
printf(\ni5%dj5%d,i,j);
getch();
if(i<=j)
{
temp=List[i];
List[i]=List[j];
List[j]=temp;
}
}
temp=List[j];
List[j]=List[lb1];
List[lb1]=temp;
if(j>lb)quicksort(List,lb,j1);
if(j<ub)quicksort(List,j+1,ub);
}
3.2.5.10 Analysis of Quick Sort
The quick sort requires the following operations:
Step
1.
2.
3.
4.

If number of elements = 1, then list is sorted and exit.

partition the list into two sub-lists around a pivot.
Place pivot in the middle of two sub-lists.
Repeat steps 1 to 3.

The best case for quick sort algorithm comes when we split the input as
evenly as possible into two sub-lists. Thus in the best case, each sub-list
would be of size n/2. Anyway, the partitioning operation would require N
comparisons to split the list of size N into the required sub-lists.
Now, for a list of size N, the number of comparisons can be defined as
follows:
NumComp(N
)

0 when N = 1

N + NumComp(sub-list1 of size N/2)

+ NumComp(sub-list2 of size N/2)

N + 2*NumComp(N/2)

The above relation is same as defined in merge sort; therefore, it can be

deduced to the following:
T(N) = O(N*log2N)
Thus, the best case time complexity of quicksort is of the order Nlog 2N or
simply n log n.
3.2.5.11 Shell Sort
It is an improvement on insertion sort. Given a list of List[N], it is
divided into k sets of N/k items each. k can assume values from a set {1,
3, 5, 19, 41}. Thus, the ith set will have the following elements:
Seti=List[i]List[i+k]List[i+2k]
Consider the list given in Figure 3.21.
For k = 5, the list is divided into the following sets:
Set 0 = 8, 7
Set 1 = 5, 19

Fig. 3.21 The sets of a list for k = 5

Fig. 3.22 The list after first pass

Set 2 = 6, 3
Set 3 = 2
Set 4 = 4
Now on each set, the insertion sort is applied resulting in the
arrangement shown in Figure 3.22.
For k = 3, the list is divided into the following sets:
Set 0 = 7, 2, 19
Set 1 = 5, 4, 6
Set 2 = 3, 8
Now on each set, the insertion sort is applied resulting into the
arrangement shown in Figure 3.23.
For k = 1, the list is represented by the following set:
Set 0 = 2, 4, 3, 7, 5, 8, 19, 6

Application of insertion sort on the above set results in the arrangement

shown in Figure 3.24
Fig. 3.23 The list after second pass
Fig. 3.24 The list after third pass
It may be noted that the list is now sorted after the third pass. Since the
variable k takes the diminishing values5, 3, and 1; the shell sort is also
called diminishing step sort.
The algorithm for the shell sort is given below. In this algorithm, a list of
elements, stored in an array called List[N], are sorted in an ascending
order. The algorithm divides the list into sets as per the description given
above. The diminishing step values are stored in a list
called dimStep. A variable sis used that moves the insertion operation
to the next set. The variable k takes the step size fromdimStep and
moves the index i within the set from one element to another.
Algorithm shellSort()
{
Step
1.initializethesetdimSteptovalues1,3,5,
/*sortthelistbyshellsort*/
2.for(step=0;step<3;step++)
{
2.1k=dimStep[step];/*setktodiminishing
step*/
2.2s=0;/*startfromthesetofthelist*/
2.3for(i=s+k;i<size;i+=k)
{
temp=List[i];/*savetheelementfromtheset
*/
j=ik;
/*findtheplaceforinsertion*/
while((temp<List[j])&&(j>=0))
{
List[j+k]=List[j];
j=jk;
}

List[j+k]=temp;/*insertthesavedelement
atitsplace*/
s++;/*gotonextset*/
}
}
3.printthesortedlist
4.Stop
}
Example 14: Given is a list of N randomly ordered numbers. Write a
program that sorts the list in ascending order by using shell sort.
Solution: The required program uses the algorithm shellSort().
/*Thisprogramsortsagivenlistofnumbersin
ascendingorderusingShellsort*/
#include<stdio.h>
#defineN20
voidmain()
{
intList[N];
intsize;
inti,j,k,p;
inttemp,s,step;
intdimStep[]={5,3,1};/*thediminishingsteps*/
printf(\nEnterthesizeofthelist(<20));
scanf(%d,&size);
printf(\nEnterthelistonebyone);
for(i=0;i<size;i++)
{
scanf(%d,&List[i]);
}
/*sortthelistbyshellsort*/
for(step=0;step<3;step++)
{
k=dimStep[step];/*setktodiminishingstep*/
s=0;/*startfromthesetofthelist*/
for(i=s+k;i<size;i+=k)
{
temp=List[i];/*savetheelementfromtheset
*/
j=ik;

while((temp<List[j])&&(j>=0))/*findthe
placeforinsertion*/
{
List[j+k]=List[j];
j=jk;
}
List[j+k]=temp;/*insertthesavedelement
atitsplace*/
s++;/*gotonextset*/
}
}/*Printthesortedlist*/
printf(\nThesortedlistis.);
for(i=0;i<size;i++)
{
printf(%d,List[i]);
}
}
The analysis of shell sort is beyond the scope of the book.
3.2.5.12 Radix Sort
It is a non-comparison-based algorithm suitable for integer values to be
sorted. It sorts the elements digit by digit. It starts sorting the elements
according to the least significant digit of the elements. This partially
sorted list is then sorted on the basis of second least significant bit and
so on.
Note: This algorithm is suitable to be implemented through linked lists.
Therefore, discussion on radix sort is currently being postponed and it
would be dealt with later in the book.
3.3 MULTI-DIMENSIONAL ARRAYS

An array having more than one subscript is called multi-dimensional

array. C supports arrays of arbitrary dimensions. However, we will
restrict ourselves to two dimensions only. A two-dimensional array has
two subscripts. It represents two-dimensional entities such as tables,
matrices etc. Therefore, sometimes the two-dimensional arrays are also
called matrix arrays.
Consider the matrix A given in Figure 3.25. The matrix A has three rows
and four columns. It contains 12 elements. Such a matrix can be
represented by a two-dimensional array, capable of storing 12 elements
in a row-column fashion. Let MAT_A[3][4] be the required array with

two subscripts. The first subscript denotes rows of the matrix varying
from 0 to 2 whereas the second subscript denotes the columns varying
from 0 to 3.

Fig. 3.25 Matrix A

Now the matrix A of Figure 3.25 can be comfortably stored in the array
called MAT_A[3][4] as shown in Figure 3.26.
Let us consider a situation where it is desired that a value 5 be added to
all elements of the matrix A, stored in array called MAT_A. This
operation can be done by picking the element MAT_A[0][0] (i.e., zeroth
row and zeroth column) and adding the required value 5 to its contents
as shown below:

Fig. 3.26 The two-dimensional array representing the matrix A

MAT_A[0][0]=MAT_A[0][0]+5;
Now to complete the rest of the operations, there is a choice, i.e., either
move along the zeroth row (MAT_A[0][1],MAT[0][2]) and add 5
to each visited element or to move along the zeroth col. (MAT_A[1][0],
MAT[2][0]) and add 5 to each visited element. Once a row or column
is over, we can move to the next row or column in the order and continue
to travel in the same fashion.
Thus, a two-dimensional array can be travelled row by row or column
by column. The former method is called as row major order and the
later as column major order.
An algorithm for row major order of travel of the array MAT_A is given
below. This algorithm visits every element of the array MAT_A and adds
the value 5 to the contents of the visited element.
Algorithm rowMajor()
{
Step
1.forrow=0to2
2.forcol=0to3
MAT_A[row][col]5MAT_A[row][col]+5;

3.Stop
}
In the above algorithm, nested loops have been used. The outer loop is
employed to travel the rows and the inner to travel the columns within a
given row.
Example 15: Given two matrices A[]mr and B[]rn , compute the
product of two matrices such that:
C[]mn=A[]mr*B[]rn
Solution: The multiplication of two matrices requires that a row of
matrix A be multiplied to a column of matrix B to generate an element of
matrix C as shown in Figure 3.27.
Fig. 3.27 The matrix multiplication operation
A close observation of the computation shown in Figure 3.27 gives the
following relation for an element of matrix C:
C[I][J]5C[I][J]+A[I][K]B[K][J]
0 to r1

for k varying from

We would use the above relation to compute the various elements of the
matrix C. The required program is given below:
/*ThisprogrammultipliestwomatricesA&Band
storestheresultantmatrixinC*/
#include<stdio.h>
main()
{
intA[10][10],B[10][10],C[10][10];
inti,j,k;/*Thearrayindices*/
intm,n,r;/*Orderofmatrices*/
printf(\nEntertheorderofthematricesm,r,
n:);
scanf(%d%d%d,&m,&r,&n);
printf(\nEnterTheelementsofMatrixA:);
for(i=0;i<m;i++)
for(j=0;j<r;j++)
{
printf(\nEnteranelement:);

scanf(%d,&A[i][j]);
}
printf(\nEnterTheelementsofMatrixB:);
for(i=0;i<r;i++)
for(j=0;j<n;j++)
{
printf(\nEnteranelement:);
scanf(%d,&B[i][j]);
}
/*Multiplythematrices*/
for(i=0;i<m;i++)
{
for(j=0;j<n;j++)
{
C[i][j]=0;
for(k=0;k<r;k++)
{
C[i][j]=C[i][j]+A[i][k]*B[k][j];
}
}
}
printf(\nTheresultantmatrixis\n);
for(i=0;i<m;i++)
{
for(j=0;j<n;j++)
{
printf(%d,C[i][j]);
}
printf(\n);
}
}
Example 16: A nasty number is defined as a number which has at least
two pairs of integer factors such that the difference of one pair equals to
sum of the other pair. For instance, the factors of 6 are 1, 2, 3, and 6.
Now the difference of factor pair (6, 1) is equal to sum of factor pair (2,
3), i.e., 6 1 = 2 + 3. Therefore, 6 is a nasty number.
Choose appropriate data structure and write a program that displays all
the nasty numbers present in a list of numbers.

Solution: The following data structures would be used to store the list
of numbers and the list of factorpair:
1. A one-dimensional array called List to store the list of numbers.
2. A two-dimensional array called pair_of_factors to store the list of pairfactors of a given number.

For example, the list of pair factors of 24 would be stored as shown

below:
1

24

12

Finally, using a nested loop, the difference of pair factors would be

compared with the sum of all pair of factors for equality. The required
program is given below:
/*Thisprogramfindsallthenastynumbersfroma
givenlistofintegernumbers*/
#include<stdio.h>
#include<math.h>
main()
{
intLIST[20];
intpair_of_factors[50][2];
inti,j,k,size,num,diff,sum,count;
printf(\nEnterthesizeofthelist:);
scanf(%d,&size);
printf(\nEnterthelist:);
for(i=0;i<size;i++)
scanf(%d,&LIST[i]);
for(i=0;i<size;i++)
{
num=LIST[i];

count=0;
/*Computethefactors*/
pair_of_factors[0][0]=1;
pair_of_factors[0][1]=num;
for(j=2;j<=sqrt(num);j++)
{
if(num%j==0)
{
count++;
pair_of_factors[count][0]=j;
pair_of_factors[count][1]=num/j;
}
}
/*Checkfornastiness*/
for(j=0;j<=count;j++)
{
diff=pair_of_factors[j][1]pair_of_factors[j]
[0];
for(k=0;k<=count;k++)
{
sum=pair_of_factors[k][1]+
pair_of_factors[k][0];
if(diff==sum)
{
printf(\n%disaNastynumber,num);
break;
}
}/*Endofkloop*/
}/*Endofjloop*/
}/*Endofiloop*/
}/*Endofprogram*/
3.4 REPRESENTATION OF ARRAYS IN PHYSICAL MEMORY

A programmer looks upon one-dimensional and two-dimensional arrays

as lists and tables (or matrices), respectively. This is a logical view of
the programmer about arrays as data structures, capable of storing lists
or tables. However, at physical level, the main memory is RAM which is
just a linear memory. The relationship between logical and physical
views of arrays is shown in Figure 3.28.
Thus, the compiler has to map the various locations of arrays to locations
of the physical memory. In fact, for a given array index, the compiler

computes its physical load time address of physical memory location. For
example, List[5] would be loaded at physical address 5 105, in the
physical memory (see Figure 3.28).
Since physical memory is linear by construction, the address mapping is
quite a simple exercise. The compiler needs to know only the starting
address of the first element in the physical memory. The rest of the
addresses can be generated by a trivial formula. A brief discussion on
address computation is given in the subsequent sections.

Fig. 3.28 Address mapping by compiler from logical to physical

memory
3.4.1 Physical Address Computation of Elements of One-dimensional Arrays
Let us reconsider the array (LIST[N]) shown in Figure 3.28. The array
called LIST has six elements. It has been stored in physical memory
at a stating address 100. The address map is given below:
The address of 1st element

100

2nd element

100 + 1

3rd element

100 + 2

100 + 5

_
_
6th element

It is very easy to deduce from above address map that the equivalent
physical address of an Ith element of array = 100 + (I 1). The starting
address of array at physical memory (i.e., 100) is called base address
(Base) of the array. For example, the base address of array called LIST is
100.
The formula for address of Ith location can be written as given below:
Address of List[I] = Base + (I 1)

(3.5)

Eq. 3.5 is a simple case, i.e., it is assumed that every element of List
would occupy only one location of the physical memory. However if an
element of an array occupies S number of memory locations, theEq.
3.5 can be modified as as follows:
Address of List[I] = Base 1 (I 1) * S

(3.6)

where
List: is the name of the array.
Base: is starting or base address of array.
I: is the index of the element.
S: is the number of locations occupied by an element of the array.
3.4.2 Physical Address Computation of Elements of Two-dimensional Arrays
Let us reconsider the two-dimensional array called TAB[2][3] as
shown in Figure 3.29.
The array called TAB has two rows and three columns. It has been
stored in physical memory at a starting address 100. The address map is
given below:
The address of 1st row, 1st col

100

1st row, 2nd col

100 + 1

1st row, 3rd col

100 + 2

100 + 5

_
_
2nd row, 3rd col

Fig. 3.29 Address mapping by compiler from logical to physical

memory.
We know that in two-dimensional array called TAB, for every row, there
are three columns. Thus, for an element of Ith row the starting address
becomes: Base + (I 1)*3 and if the element in row is at Jth column, the
complete address becomes:
Address of TAB[I][J] = Base + (I 1)* 3 + J 1

(3.7)

To test the above row major order formula, let us calculate the address of
element represented byTAB[2][1] with Base = 100.
Given: I = 2, J = 1, Base = 100. Putting these values in Eq. 3.7, we get
Address = 100 + (2 1)*3 + 1 1 = 100 + 3 + 0 = 103.
From Figure 3.29, we find that the address value is correct.
Now we can generalize the formula of ? Eq. 3.7 for an array of size [M]
[N] in the following form:
Address of [I][J] the element = BASE + (I 1)*N + (J
1)
(3.8)
The formula given above is a simple case, i.e., it is assumed that every
element of TAB would occupy only one location of the physical memory.
However if an element of an array occupies S number of memory
locations, the row major order formula (3.8) can be modified as given
below:
Address of TAB [I][J] the element = BASE + ((I 1)*N + (J
1))*S
(3.9)

where
TAB: is the name of the array.
Base: is starting or base address of array.
I: is the row index of the element.
J: is the column index of the element.
N: is number of columns present in a row.
S: is the number of locations occupied by an element of the array.
Similarly, the formula for column major order is as follows:
Address of TAB [I][J] the element = BASE + ((J 1)*M + (I
1))*S
(3.10)
where
TAB: is the name of the array.
Base: is starting or base address of array.
I: is the row index of the element.
J: is the column index of the element.
M: is number of rows present in a column.
S: is the number of locations occupied by an element of the array.
Example 17: TAB is a two-dimensional array with five rows and three
columns. Each element occupies one memory location. If TAB[1][1]
begins at address 500, find the location of TAB [4][2] for row major
order of storage.
Solution:
Given
base

500, I = 4, J = 2, M = 5, N = 3. Applying formula 3.9,

we get:

Address

Base + ((I 1)*N + (J 1))*S

500 + ((4 1)*3 + (2 1))*1

500 + (9 + 1)

510.

Example 18: MAT is a two-dimensional array with ten rows and five
columns. Each element is stored in two memory locations. If MAT[1]

[1] begins at address 200, find the location of MAT[3][4] for row
major order of storage.
Solution:
Given
base

200, I = 3, J = 4, M = 10, N = 5. Applying formula 3.9,

we get:

Address

Base + ((I 1)*N + (J 1))*S

200 + ((3 1)*5 + (4 1))*2

200 + (10 + 3)*2

226.

Example 19: An array A[10][20] is stored in the memory with each

element requiring two memory locations. If the base address of the array
in the memory is 400, determine the location of array element A[8][3]
when array is stored as column major order.
Solution:
Given
base

400, I = 8, J = 3, M = 10, N = 20. Applying formula

3.10, we get:

Address

Base + ((J 1)*M + (I 1))*S

400 + ((3 1)*10 + (8 1))*2

400 + (20 + 7)*2

454.
3.5 APPLICATIONS OF ARRAYS

The arrays are used data structures which are very widely in numerous
applications. Some of the popular applications are:
3.5.1 Polynomial Representation and Operations
An expression of the form f(x) = anxn + an-1 xn-1+ an-2 xn-2 + ..+ a0 is
called a polynomial

where an 0.
anxn , an-1 xn-1 a1 x1, a0 are called terms.
an, an-1, a1, a0 are called coefficient.
xn, xn-1, xn-2are called exponents,
Now, a polynomial can be considered as a list comprising coefficients
and exponents as shown below:
F(x) = {an, xn , an-1, xn-1, a1, x1, a0, x0
For example, the polynomial 6x5 + 8x2 + 5 can be represented as the list
shown below:
Polynomial 5 {6, 5, 8, 2, 5, 0}
The above list can be very easily implemented using a one-dimensional
array, say Pol[] as shown inFigure 3.30 where every alternate location
contains coefficient and exponent.
The above representation can be modified to incorporate number of
terms present in a polynomial by reserving the zeroth location of the
array for this purpose. The modified representation is shown inFigure
3.31.

Fig. 3.30 Array representation of a polynomial

It may be noted from the above discussion that a general polynomial can
be represented in an array with descending order of its degree of terms.
Therefore, the operations such as addition, multiplication, and division
on polynomials can be easily carried out.

Fig. 3.31 Array representation of a polynomial with zeroth place

reserved for number of terms
Example 20: Write a program that reads two
polynomials pol1 and pol2, adds them and gives a third
polynomial pol3 such that pol3 = pol1 + pol2.
Solution: A close look at the array representation of polynomials
indicates, that two polynomials can be added by merging them using

algorithm merge() of Section 3.2.5.7. The only change would be that

when the exponents from both the polynomials are same, then the
coefficients of both the terms would be added.
The required program is given below:
/*Thisprogramaddstwopolynomialspol1andpol2to
givethirdpolynomialcalledpol3.Itmergesthetwo
polynomialswiththehelpofAlgorithmmerge()with
minormodifications*/
#include<stdio.h>
voidreadPol(intpol[],intterms);
voiddispPol(intpol[]);
voidmain()
{
intpol1[11];
intpol2[11];
intpol3[21];
inti,j,k;
intterms;
intm,n;/*sizeofpolynomials*/
printf(\nEnterthenumberoftermsinPol1);
scanf(%d,&terms);
m=terms*2+1;
readPol(pol1,terms);
printf(\nEnterthenumberoftermsinPol2);
scanf(%d,&terms);
n=terms*2+1;
readPol(pol2,terms);
/*Addthepolynomialsbymergingtheirterms*/
i=j=k=2;
while(i<m&&j<n)
{
if(pol1[i]>pol2[j])
{
pol3[k]=pol1[i];
pol3[k1]=pol1[i1];
i=i+2;
k=k+2;
}
else
if(pol2[j]>pol1[i])

{
pol3[k]=pol2[j];
pol3[k1]=pol2[j1];
j=j+2;
k=k+2;
}
else
{
pol3[k1]=pol1[i1]+pol2[j1];
pol3[k]=pol1[i];
i=i+2;
j=j+2;
k=k+2;
}
}
/*copytheremainingPol1*/
while(i<m)
{
pol3[k]=pol1[i];
pol3[k1]=pol1[i1];
i=i+2;
k=k+2;
}
/*copytheremainingPol2*/
while(j<n)
{
pol3[k]=pol2[j];
pol3[k1]=pol2[j1];
j=j+2;
k=k+2;
}
pol3[0]=(k1)/2;
/*Displaythefinalpolynomial*/
printf(\nFinalPol=);
dispPol(pol3);
printf(\b);
}
voidreadPol(intpol[],intterms)
{
inti;
pol[0]=terms;

i=1;
printf(\nEntertheterms(coefexp)indecreasing
orderofdegree);
while(terms>0)
{
scanf(%d%d,&pol[i],&pol[i+1]);
i=i+2;
terms;
}
}
voiddispPol(intpol[])
{
intsize;
inti;
size=pol[0]*2+1;
printf(\n);
for(i=1;i<size;i+52)
{
printf(%d*x^%d+,pol[i],pol[i+1]);
}
}
We can also choose other data structures to represent polynomials. For
instance, a term can be represented by struct construct of C as shown
below:
structterm
{
intcoef;
intexp;
};
and the polynomial called pol of ten terms can be represented by an
array of structures of type term as shown below:
struct term pol1[10];
Now the algorithm merge() can be applied to add two such polynomials
with above chosen data structures. It is left as an exercise for the readers
to write that program.

3.5.2 Sparse Matrix Representation

A matrix is called sparse when it contains a large number of zero entries.
Consider the matrix given inFigure 3.32.
It may be noted that the matrix MatA of order 6 8 has 42 zero entries
out of total 48 entries. Thus, it has enough number of zero entries that
prompts us to look for an alternate representation that could save the
memory space occupied by zero entries. A popular representation of a
sparse matrix is given below.

Fig. 3.32 A matrix of order 6 * 8

A sparse matrix can be very easily represented by a list of ordered pairs.
An ordered pair has the following format
(Row,Col,E)
Where
Row:

istherownumber

Col

isthecolumnnumber

:
E:

isthenonzeroentryatthegivenrow
andcol.

Lets us now apply the above representation to MatA to obtain the list of
ordered pairs as given below:
0,4,2
1,1,3
1,6,5
3,3,7
4,7,4
5,0,8
The above representation is also called a condensed matrix. It is using
comparatively very less space which will become even prominent for
large matrices, say of the order of 500 500 or more.

The proposed list of ordered pairs can be modeled as a two dimensional

matrix of order of n + 1 3. Where n is the number of nonzero elements
present in a sparse matrix. However, the 0th ordered pair can be suitably
used to store the order of the sparse matrix and the number of nonzero
elements present in the matrix as shown in Figure 3.33.

Fig. 3.33 The condensed representation

Example 21: Write a program that reads a sparse matrix of
order m*n and stores it into the condensed representation as discussed
above i.e. a two dimensional array of size n+1*3. Where n is number
of nonzero elements. Print the contents of the condensed matrix.
Solution: The required program is given below:
/*Thisprogramreadsasparsematrixoforderm*n
andstoresitintoacondensedmatrixformat*/
#include<stdio.h>
#include<conio.h>
main()
{
intmatA[7][3];
inti,j,k;
intm,n,element;
clrscr();
printf(\nEntertheorder(m,n)ofthesparse
matrix);
scanf(%d%d,&m,&n);
matA[0][0]=m;/*no.ofrows*/
matA[0][1]=n;/*no.ofcols*/
k=1;/*pointto1stpositionofsparsemat*/
printf(\nEntertheelementsofthematrixinrow
majororder);
for(i=0;i<m;i++)
for(j=0;j<n;j++)
{
scanf(%d,&element);
if(element!=0)

{/*storenonzeroelementintocondensed
matrix*/
matA[k][0]=i;
matA[k][1]=j;
matA[k][2]=element;
k++;
}
}
matA[0][2]=k1;/*recordno.ofnonzero
elements*/
printf(\nThecondensedmatrixis);
for(i=1;i<k;i++)
printf(\n%d%d%d,matA[i][0],matA[i]
[1],matA[i][2]);
}
Note: No space has been provided for the sparse matrix. The elements
have been read one by one and the nonzero elements have been stored
into the condensed matrix.
A sample run of the program is given below:

3.5.2.1 Sparse Matrix Addition

Two sparse matrices can be added in the same way as the polynomials
were added in previous section. Consider the three sparse matrices Mat1,
Mat2, and Mat3 given in Figure 3.34.
where Mat3 = Mat1 + Mat2

Fig. 3.34 The addition of matrix

The equivalent condensed form of Mat1, Mat2, Mat3 is given in Figure
3.35.

Fig. 3.35 The equivalent condensed representation

It may be noted that the final matrix Mat3, is nothing but a merger of
two ordered lists Mat1 and Mat2 in a manner similar to polynomials. The
algorithm merge() of section 3.2.5.7 can be very easily applied. The
only change would be that when the row number and column number
from both the condensed matrices are same then the elements of both
the matrices would be added. For example, merger of (1 1 3) and (1 1 6)
would result into the element (1 1 9).
Example 22: Write a program that reads two sparse matrices Mat1 and
Mat2 of order m*n, and represents them in condensed form. The
condensed matrices are added to give a third condensed matrix called
Mat3. It prints Mat3 in row major order.
Solution: We would use following three functions:
readMat(): It reads a sparse matrix and stores it into a condensed form
addsparsMat(): It takes two condensed matrices and adds them into a
third condensed matrix
printMat(): It prints a condensed matrix
The required program is given below:
/*Thisprogramaddstwosparsematricesoforderm*n
*/
#include<stdio.h>
#include<conio.h>
voidreadMat(intmatA[][],intm,intn);
voidprintMat(intmatA[][]);
voidaddsparsMat(intmat1[][],intmat2[][],int
mat3[][]);
voidmain()
{

intmat1[10][3],mat2[10][3],mat3[20][3];
intm,n;
clrscr();
printf(\nEntertheorderofthesparsemat1:m,
n);
scanf(%d%d,&m,&n);
readMat(mat1,m,n);
printMat(mat1);
printf(\nEntertheorderofthesparsemat2:m,
n);
scanf(%d%d,&m,&n);
readMat(mat2,m,n);
printMat(mat2);
addsparsMat(mat1,mat2,mat3);
printf(\nThefinalcondensedMatrix);
printMat(mat3);
}
voidreadMat(intmatA[10][3],intm,intn)
{
inti,j,k;
intelement;
matA[0][0]=m;/*no.ofrows*/
matA[0][1]=n;/*no.ofcols*/
k=1;/*pointto1stpositionofsparsemat*/
printf(\nEntertheelementsofthematrixinrow
majororder);
for(i=0;i<m;i++)
for(j=0;j<n;j++)
{
scanf(%d,&element);
if(element!=0)
{
matA[k][0]=i;
matA[k][1]=j;
matA[k][2]=element;
k++;
}
}
matA[0][2]=k1;/*recordno.ofnonzeroelements
*/
}

voidprintMat(intmatA[10][3])
{
inti,size;
size=matA[0][2];
for(i=0;i<=size;i++)
printf(\n%d%d%d,matA[i][0],matA[i]
[1],matA[i][2]);
}
voidaddsparsMat(intmat1[10][3],intmat2[10][3],
intmat3[20][3])
{
inti,j,k;
ints1,s2;/*No.ofelementsinmat1andmat2*/
s1=mat1[0][2];
s2=mat2[0][2];
i=j=k=0;
while(i<=s1&&j<=s2)
{
if(mat1[i][0]<mat2[j][0])/*rowofmat1<mat2*/
{
mat3[k][0]=mat1[i][0];
mat3[k][1]=mat1[i][1];
mat3[k][2]=mat1[i][2];
k++;i++;
}
else
if(mat1[i][0]>mat2[j][0])/*rowofmat1>mat2
*/
{
mat3[k][0]=mat2[j][0];
mat3[k][1]=mat2[j][1];
mat3[k][2]=mat2[j][2];
k++;j++;
}
else
{/*rowofmat1=mat2*/
if(mat1[i][1]<mat2[j][1])/*colofmat1<
mat2*/
{
mat3[k][0]=mat1[i][0];
mat3[k][1]=mat1[i][1];

mat3[k][2]=mat1[i][2];
k++;i++;
}
else
if(mat1[i][1]>mat2[j][1])/*colofmat1>
mat2*/
{
mat3[k][0]=mat2[j][0];
mat3[k][1]=mat2[j][1];
mat3[k][2]=mat2[j][2];
k++;j++;
}
else
{/*colofmat1=mat2*/
mat3[k][0]=mat2[i][0];
mat3[k][1]=mat2[i][1];
mat3[k][2]=mat1[i][2]+mat2[j][2];
k++;i++;j++;
}
}
}
while(i<=s1)/*copyrestoftheelementsof
mat1*/
{
mat3[k][0]=mat1[i][0];
mat3[k][1]=mat1[i][1];
mat3[k][2]=mat1[i][2];
k++;i++;
}
while(j<=s2)/*copyrestoftheelementsofmat2
*/
{
mat3[k][0]=mat2[j][0];
mat3[k][1]=mat2[j][1];
mat3[k][2]=mat1[j][2];
k++;j++;
}
mat3[0][2]=k1;
}
A sample output of the program is given below:

3.5.2.2 Sparse Matrix Transpose

The transpose of a matrix of order m*n is defined as a matrix n
*m that is formed by interchanging the rows and columns of the matrix.
In fact, the transpose of the matrix is obtained by exchanging the
element at (i,j) th position in the matrix t with the element at (j,
i)th place as shown in Figure 3.36.

Fig. 3.36 Transpose of a sparse (condensed) matrix

It may be noted that the number of rows and number of columns have
also been exchanged.
Example 23: Write a program that produces a transpose of a given
sparse matrix of order m*n.
Solution: The required program is given below
/*ThisprogramproducesaTransposeofasparse
matrix*/
#include<stdio.h>
#include<conio.h>
voidreadMat(intmatA[][],intm,intn);

voidprintMat(intmatA[][]);
voidtransMat(intmat[][],inttrans[][]);
voidmain()
{
intmat[10][3],trans[10][3];
intm,n;
clrscr();
printf(\nEntertheorderofthesparsemat:m,
n);
scanf(%d%d,&m,&n);
readMat(mat,m,n);
printMat(mat);
transMat(mat,trans);
printf(\nThetransposedcondensedMatrix);
printMat(trans);
}
voidreadMat(intmatA[10][3],intm,intn)
{
inti,j,k;
intelement;
matA[0][0]=m;/*no.ofrows*/
matA[0][1]=n;/*no.ofcols*/
k=1;/*pointto1stpositionofsparsemat*/
printf(\nEntertheelementsofthematrixinrow
majororder);
for(i=0;i<m;i++)
for(j=0;j<n;j++)
{
scanf(%d,&element);
if(element!=0)
{
matA[k][0]=i;
matA[k][1]=j;
matA[k][2]=element;
k++;
}
}
matA[0][2]=k1;/*recordno.ofnonzeroelements
*/
}
voidprintMat(intmatA[10][3])

{
inti,size;
size=matA[0][2];
for(i=0;i<=size;i++)
printf(\n%d%d%d,matA[i][0],matA[i][1],matA[i]
[2]);
}
voidtransMat(intmat[10][3],inttrans[10][3])
{
inti,j,small,size,pos,temp;
size=mat[0][2];
trans[0][0]=mat[0][1];
trans[0][1]=mat[0][0];
trans[0][2]=mat[0][2];
for(i=1;i<=size;i++)
{
trans[i][0]=mat[i][0];
trans[i][1]=mat[i][1];
trans[i][2]=mat[i][2];
}
for(i=1;i<size;i++)
{
small=trans[i][1];pos=i;
for(j=i+1;j<=size;j++)
{
if(small>trans[j][1])
{
small=trans[j][1];
pos=j;
}
else
if(small==trans[j][1]&&trans[i][0]>trans[j]
[0])
{
small=trans[j][1];
pos=j;
}
}
temp=trans[i][0];trans[i][0]=trans[pos]
[0];trans[pos][0]=temp;

temp=trans[i][1];trans[i][1]=trans[pos]
[1];trans[pos][1]=temp;
temp=trans[i][2];trans[i][2]=trans[pos]
[2];trans[pos][2]=temp;
temp=trans[i][0];trans[i][0]=trans[i][1];
trans[i][1]=temp;
}
}
A sample output is given below:

EXERCISES
1. Define the terms: array, subscript, subscripted variables and strings.
2. Write suitable array declaration for the following:

1.
2.
3.
4.
2
3
4
5
6
7
8

100 items of integer type.

A string of 25 characters.
A matrix of order 5 4.
Can an array be initialized? If yes, how?

What is the purpose of initializing an array?

Can array be initialized at the time of declaration?
How can a string be stored in an array?
Write a program that finds the largest and smallest elements in an array.
Write a program that removes duplicate from a list.
Write a program that removes an element from an array.
Write a program that searches an array SOME of float type for a value less
that x. The program should give appropriate message depending upon the
outcome of the search.
9 Write a program which searches an ordered array SAL of integer type for a
value X.

10 Write a program that gives the sum of positive elements of an array LIST of
integer type.
11 Write a program that reads two matrices A and B of order m n and
compute the following:
C=A+B
12 What happens to an array if it is assigned with less number of elements at
the time of initialization?
13 Write a program that computes the sum of diagonal elements of a square
matrix.
14 Write a program that takes a sparse matrix A and finds its transpose AT and
displays it.
15 Let A be an N N square matrix array. Write algorithms for the following:

1. Find the number of non-zero elements in A.

2. Find the product of the diagonal elements (a11, a22, a33,
ann).
2 Find out the complexity of binary search. Compare its time complexity with
linear search.
3 What is meant by sorting? What are the types of sorting?
4 What is the difference between bubble sort and selection sort?
5 What is meant by quick sort? What are the advantages/disadvantages of
quick sort over merge sort?
6 Explain how the selection of pivot plays major role in efficiency of quick sort
algorithm and determine the complexity of the quick sort.
7 Trace the steps of insertion sort for the list of numbers: 12, 19, 33, 26, 29,
35, 22. Compute the total number of comparisons made.
8 What is shell sort? Define the worst case analysis of shell sort.
9 What is merge sort? Write algorithm for merge sort and derive its run time
complexity.
10 How the two-dimensional arrays are represented in memory? Also, obtain
the formula for calculating the address of any element stored in the array, in
column major order.
11 Write a program that sorts a given list of numbers using bubble sort.
12 Write a program that sorts a given list of numbers using selection sort.
13 Derive the time complexity of selection sort.
14 Derive the time complexity of insertion sort.
15 Write a program that sorts a given list of numbers using quick sort.
16 Write a program that sorts a given list of numbers using merge sort.

4
Stacks and Queues
CHAPTER OUTLINE

4.1 Stacks
4.2 Applications of Stacks
4.3 Queues
4.1 STACKS

We are familiar with arrays and the operations performed on them. The
various elements can be accessed from any position in a list, stored in an
array. Some practical applications require that the additions or deletions
be done only at one end. For example, the packets of wheat floor received
in a super bazaar are put on top of another in the order of their arrival.
Now, a customer can remove only the packet currently lying on the top,
i.e., the last packet arrived in the bazaar is the first one to leave. The
second customer would get the second last packet and so on. In the mean
time, if more packets arrive then they will be added on the current top as
shown in Figure 4.1.

Fig. 4.1 The stack

It may be noted that in this kind of arrangement, all additions and
deletions are made only from one end called top. In general, this activity
is called stacking and the structure that allows deletions and additions
from one end (top) is called a stack. It is also called LIFO (Last In First
Out) data structure.

In a restaurant, the dish washing unit maintains the washed plates in a

spring-loaded container. The newly washed plate is pushed on the top of
the container and the customer may pop a plate from the top only.
Therefore, the addition and deletion operations from a stack are
accordingly also called Push and Pop, respectively as shown in Figure
4.2.

Fig. 4.2 The stack operations

Similarly, the nested calls to functions and procedures are also handled
in LIFO fashion in the sense that control from last called procedure is
handled first.
A stack can be more precisely defined as a linear collection of items
which allows addition and deletion operations only from one end called
top.
The push operation adds an item onto the top of the stack whereas the
pop operation removes an element from the current top of the stack. The
size of the stack depends upon the number of items present in the stack.
With the addition and deletion of items, the size of stack accordingly
increases or decreases. Therefore, stack is also called a dynamic data
structure with a capacity to enlarge or shrink.
4.1.1 Stack Operations
As already discussed, there are two main operations (pushand
pop) associated with a stack. The push operation adds an item on the
top of the stack whereas the pop operation removes an item from the
top of the stack. Let us consider the stack shown in Figure 4.3 (a).
Currently the Top, a variable, is pointing to the item D. Let us now push
an item E on the stack. The Top is incremented to next location and
the item is added at the pointed location as shown in Figure 4.3(b).

Fig. 4.3 Push operation on stack

It may be noted that the variable called Top keeps track of the location
where addition can be done. If there is a bound on the size of
the stack (sayN) then as soon as the Top becomes equal to N,
thestack is said to be full.
Let us now apply pop operation three times in succession and see the
results. The following three items will be removed from
the stack in LIFO fashion:
E D C
The variable top is also decremented accordingly after each Pop
operation as shown in Figure 4.4.

Fig. 4.4 Pop operations on the stack

It may be noted that if more pop operations are done on the stack then
a stage will come when there will be no items left on the stack. This
condition is called stack empty.
The algorithms for push and pop operations are given below. These
algorithms use an array calledStack[N] to represent
a stack of N locations. A variable called Top keeps track of the top of
thestack, i.e., the location where additions and deletions are made.
Another variable called item is used to store the item to be pushed or
popped from the stack.
The algorithm for push is given below:
Algorithm Push ( )
{
Step
1. If(Top>=N)then{prompt(StackFull);exit}
2. Top=Top+1;
3. Stack[Top]=item;
4. Stop
}
It may be noted that the item is added to
the stack (i.e., push operation) provided the stack is not already full,
i.e., at step1, it is checked whether or not the stack has room for another
item.

The algorithm for pop is given below:

Algorithm Pop()
{
Step
1.if(Top0)henprompt(StackEmpty);
else
Item=Stack[Top];
Top=Top1;
2.stop
}
It may be noted that in the above algorithm, before popping an item
from the top, it is checked at step 1 to see if there is at least one item on
the stack that can be removed.
The performance of the stack data structure for n elements is:
o The space complexity is O(n).
o The time complexity of each operation is O(1).
o The size of the stack, implemented using array, must be defined in
advance, i.e., a priori. Moreover, the size cannot be changed.

Example 1: Write a program in C that uses the following menu to

simulate the stack operations on items of int type.

Enter your choice:

Solution: The above defined algorithms Push ( ) and Pop ( ) would be
employed to write the program. The menu would be created
using printf ( ) statements and displayed within a do-while loop. In
fact,a do-while loop is the most appropriate loop for display and
manipulations of menu items.
Note:
o The Push ( ) function would return 1 if the operation is successful and 0
otherwise, i.e., 0 indicates that the stack was full.

o The Pop ( ) function would return the popped value if the operation is
successful and 9999 otherwise, i.e., the unusual value 9999 indicates that
the stack was empty.

The required program is given below:

/*Thisprogramsimulatesstackoperations*/
#include<stdio.h>
#include<conio.h>
intpush(intstak[],int*top,intval,intsize);
intpop(intstak[],int*top);
voiddisplay(intstack[],inttop);
voidmain()
{
intstack[20];/*thestackdeclaration*/
inttop=1;/*initiallythestackisempty*/
intval;
intsize;
intchoice;
intresult;
printf(\nEnterthesizeofthestack);
scanf(%d,&size);
size;/*adjustedforindex=0*/
/*createmenu*/
do
{
clrscr();
printf(\nMenuStackOperations);
printf(\nPush1);
printf(\nPop2);
printf(\nDisplaystack3);
printf(\nQuit4);
printf(\nEnteryourchoice:);
scanf(%d,&choice);
switch(choice)
{
case1:printf(\nEnterthevaluetobe
pushed);
scanf(%d,&val);
result=push(stack,&top,val,size);
if(result==0)

printf(\nThestackfull);
break;
case2:result=pop(stack,&top);
if(result==9999)
printf(\nThestackisempty);
else
printf(\nThepoppedvalue=%d,result);
break;
case3:display(stack,top);
break;
}
printf(\n\nPressanykeytocontinue);
getch();
}
while(choice!=4);
}
intpush(intstack[],int*top,intval,intsize)
{
if(*top>=size)
return0;/*thestackisFull*/
else
{
*top=*top+1;
stack[*top]=val;
return1;
}
}
intpop(intstack[],int*top)
{intval;
if(*top<0)
return9999;/*thestackisempty*/
else
{
val=stack[*top];
*top=*top1;
returnval;
}
}
voiddisplay(intstack[],inttop)

{
inti;
printf(\nThecontentsofstackare:);
for(i=top;i>=0;i)
printf(%d,stack[i]);
}
4.2 APPLICATIONS OF STACKS

The stacks are used in numerous applications and some of them are as
follows:
o
o
o
o
o

Arithmetic expression evaluation

Undo operation of a document editor or a similar environment
Implementation of recursive procedures
Backtracking
Keeping track of page-visited history of a web user

4.2.1 Arithmetic expressions

An arithmetic expression is a combination of variables and/or constants
connected by arithmetic operators and parenthesis. The valid arithmetic
operators are given in Table 4.1.
Table 4.1 Arithmetic operators
Symbol

Stands for

Exponentiation

X^8

Division

X/Y

Multiplication

X*Y

Addition

X+Y

Subtraction

XY

The rules for an arithmetic expression are:

o A signed or unsigned variable or constant is an expression.
o An expression connected by an arithmetic operator to a variable or constant
is an expression.
o Two expressions connected by an arithmetic operator is also an expression.
o Two arithmetic expressions should not occur in continuation.

Examples of some valid arithmetic expressions are:

o
o
o
o

Sum + 10
Total 100
Total/Rate*Interest
X^Y B*C

Examples of some invalid arithmetic expressions are:

o X*/Y (two operators in continuation)
o X Y (expression without an operator)

The order of evaluation, known as operator precedence, is carried out

according to the priority of the arithmetic operators given in Table 4.2.
Table 4.2 Operator precedence
S. no.

Operator(s)

^, exponentiation

/ *, division, multiplication

+ , addition, subtraction

Note:
o The operators of same priority like (*, /) or (+, ) are performed from left to
right.
o The operators contained within parentheses are evaluated first. When the
parentheses are nested, the innermost is evaluated first and so on.
o The exponential operators are evaluated from right to left. For example, in
the following expression,

A^B^C
The sub-expressions are evaluated from right to left as shown in Figure
4.5.
The arithmetic expression, as discussed above, is called
an infix expression wherein the operator is placed between two
operands. The evaluation takes place according to priorities assigned to
the operators (see Table 4.2). However, to overrule the priority of an
operator, parenthesis is used. For example, the two infix expressions
given below are entirely different because in Expression 2, the priority of
the division operator / has been overruled by the embedded
parenthesis.

Fig. 4.5 Evaluation of exponential operator

1. X Y/Z
2. ( X Y )/Z

The order of evaluation of above infix expressions is shown in Figure

4.6.

Fig. 4.6 Evaluation order of infix expressions

Thus, parenthesis play an important role in the evaluation of
an infix expression and the order of evaluation depends upon the
placement of parentheses and the operator precedence. Therefore,
theinfix representation of arithmetic expressions becomes inefficient
from the compilation point of view. The reason being that to find the
sub-expression with highest priority, at each step, repeated scanning of
the expression is required from left to right.
A Polish logician J. Lukasicwicz introduced a notation which permits
arithmetic expressions without parentheses. The absence of embedded
parentheses allows simpler compiler interpretation, translation, and
execution of arithmetic expressions. This notation is also called Polish
notation. It appears in two forms: prefix and postfix forms.
4.2.1.1 Prefix Expression (Polish Notation)
In the prefix form, an operator is placed before its operands, i.e., the
operators are prefixed to their operands. For example,
the infix expression A + B is written as +AB in prefix form.
Examples of some valid prefix expressions with their
equivalent infix expressions are given in Table 4.3.
Table 4.3 Some prefix expressions
prefix expression

equivalent infix expression

+AB

A+B

*+ABC

(A + B)*C

+/A + BC*D EF

A/(B + C) + D*(E F)

4.2.1.2 Postfix Expression (Reverse Polish Expression)

In the postfix expression, the operator is placed after its operands, i.e.,
the expression uses suffix orpostfix operators. For example,
the infix expression A + B is written as A B + in postfix form.
Examples of some valid postfix expressions with their
equivalent infix expressions are given inTable 4.4.
Table 4.4 Some postfix expressions
prefix expression

equivalent infix expression

AB+

A+B

AB + C*

(A + B)*C

ABC +/DEF *+

A/(B + C) + D*(E F)

It may be noted that in Polish and reverse Polish notations, there are no
parentheses and, therefore, the order of evaluation will be determined by
the positions of the operators and related operands in the expression.
A critical look at the postfix expressions indicates that such
representation is excellent from the execution point of view. The reason
being that while execution, no consideration has to be given to the
priority of the operators rather the placement of operators decides the
order of execution. Moreover, it is free from parentheses. Same is true
for prefix expressions but in this book, only the postfixexpressions
would be considered.
Thus, there is a need for conversion of infix to postfix expression.
The subsequent section discusses two methods to convert
the infix expressions to postfix expressions.
4.2.1.3 Conversion of Infix Expression to Postfix Expression
An infix expression can be converted into postfix by two methods:
parenthesis method and stackmethod.
(1) Parenthesis method The following steps are used to convert
an infix expression to postfixexpression by parenthesis method:
1. Fully parenthesize the infix expression.

2. Replace the right hand parenthesis by its corresponding embedded

operator.
3. Remove the left hand parenthesis. The resultant expression is
in postfix notation.

Consider the infix expression A/(B + C) + D*(E F). Let us apply the
above steps to convert this expression to postfix notation. The
operations are shown in Figure 4.7.

Fig. 4.7 Conversion of infix to postfix expressions

The method discussed above is rather inefficient because it requires
following two passes over the expression:
1. The first pass to fully parenthesize the expression.
2. The second pass to replace the right hand parentheses by their
corresponding embedded operators.

It may be noted that the infix expression within innermost

parentheses should be the first to be converted into postfix before
expressions in outermost parentheses are picked. This successive
elimination of parentheses is done until whole of the expression is
converted into postfix. This LIFOnature of the 2nd pass suggests that
a stack can be used for the conversion process.
(2) The stack method In this method, two priorities are assigned to
the operators: instack priority and incoming priority as shown in Table
4.5.
Table 4.5 Priorities assigned to operator in the stack method

It may be noted that the left hand parenthesis ( has the highest
incoming priority and least instack priority.
Let us assume that an infix expression is terminated by the character
;. We shall call operators and operands of an expression by a general
name element. The following broader level algorithm convert ( ) can be
used to convert an infix expression to postfix expression.
Algorithm convert( )
{
Step
1.Scantheexpressionfromlefttoright.
2.Iftheelementisanoperand,thenstorethe
elementintothetargetarea.
3.Iftheelementisanoperator,thenpushitonto
astackwithfollowingrules:
3.1Whiletheincomingpriorityislessthanor
equaltoinstackpriority,poptheoperatorsand
storeintothetargetarea.
3.2Iftheelementisrighthandparenthesis),
thenpopalltheelementsfromthestacktillleft
handparenthesis(ispoppedout.Thepopped
elementsarestoredinthetargetarea.
4.Iftheelementis;,thenpopalltheelements
tillthestackisemptyandstorethemintothe
targetarea.
5.Stop
}
Example 2: Write a program that uses a stack to convert
an infix expression to a postfixexpression.
Solution: We would employ the algorithm convert( ), given above.
The following data structures would be used:
The list of allowed binary operators along with their icp and isp is
stored in an array of structures called opTab[]. An array
called stack of type operator A is the main data structure. match ( )
function is being used to find out as to whether a given element is an
operator or an operand depending upon the result (res) of match
operation. The required program is given below:

/*Thisprogramconvertsaninfixexpressionto
apostfixexpression*/
#include<stdio.h>
#include<conio.h>
structoperator{/*operatordeclaration*/
charopName;
intisp;
inticp;
};
intpush(structoperatorstak[],int*top,struct
operatorval,
intsize);
structoperatorpop(structoperatorstak[],int
*top);
voiddisplay(structoperatorstack[],inttop);
intmatch(structoperatoropTab[8],charelement,
struct
operator*op);
voidmain()
{
charinfix[20];
chartarget[20];
structoperatorstack[20];/*thestackdeclaration
*/
inttop=1;/*initiallythestackisempty*/
intres;
charval;
intsize;
intpos;
inti;
structoperatorop,opTemp;
structoperatoropTab[8]={(,0,6,/*The
operatorsinformation*/
),0,0,
^,4,5,
*,3,3,
/,3,3,
+,2,2,
,2,2,
;,0,1,
};

printf(\nEnterthesizeoftheinfixexpression);
scanf(%d,&size);size;
printf(\nEnterthetermsofinfixexpressionone
byone);
for(i=0;i<=size;i++)
{fflush(stdin);/*flushtheinputbuffer*/
scanf(%c,&infix[i]);
}
pos=0;/*positionintargetexpression*/
for(i=0;i<=size;i++)
{
res=match(opTab,infix[i],&op);/*find
whetheroperator/operand*/
if(res==0)
{target[pos]=infix[i];/*storeintotarget
expression*/
pos++;
}
else
{if(top<0)
push(stack,&top,op,size);/*firsttimePush
*/
else
{opTemp.opName=#;/*placeanyvaluetoopTemp
*/
if(op.opName==))
{while(opTemp.opName!=()
{
opTemp=pop(stack,&top);
if(opTemp.opName!=()/*omit(*/
{target[pos]=opTemp.opName;
pos++;
}
}
}
else
{while(stack[top].isp>=op.icp&&top>=0)
{
opTemp=pop(stack,&top);
target[pos]=opTemp.opName;
pos++;

}
push(stack,&top,op,size);
}
}
}
}
/*printthepostfixexpression*/
printf(\nThepostfixexpressionis:);
for(i=0;i<pos;i++)
{
printf(%c,target[i]);
}
}
intpush(structoperatorstack[],int*top,struct
operator
val,intsize)
{
if(*top>=size)
return0;/*thestackisfull*/
else
{
*top=*top+1;
stack[*top]=val;
return1;
}
}
structoperatorpop(structoperatorstack[],int
*top)
{structoperatorval;
if(*top<0)
{val.opName=#;
returnval;/*thestackisempty*/
}
else
{
val=stack[*top];
*top=*top1;
returnval;
}
}
voiddisplay(structoperatorstack[],inttop)

{
inti;
printf(\nThecontentsofstackare:);
for(i=top;i>=0;i)
printf(%c,stack[i].opName);
}
intmatch(structoperatoropTab[8],charelement,
structoperator*op)
{
inti;
for(i=0;i<8;i++)
{
if(opTab[i].opName==element)
{
*op=opTab[i];
return1;
}
}
return0;
}
For following given infix expressions, the above program produces the
correct output as shown below:
Infix expression

Postfix expression

(1) A ^ B ^ C + D;

ABC^^D+

(2) (A B)*( C/D ) + E;

AB CD/* E +

Example 3: Use a stack to convert the following infix arithmetic

expression into a postfixexpression. Show the changing status of
the stack in tabular form.
A + B*C
Solution: We would use the algorithm convert(). The simulation of
this algorithm for the above expression is given in Table 4.6.
The output postfix expression is ABC* +

Example 4: Use a stack to convert the following infix arithmetic

expression into a postfixexpression. Show the changing status of
the stack in tabular form.
(A B)*(C/D) + E
Table 4.6 Conversion of A + B*C into its postfix form

Solution: We would use the algorithm convert( ). The simulation of

this algorithm for the above expression is given in Table 4.7.
The output postfix expression is AB CD/* E+
Table 4.7 Conversion of (A - B)*(C/D) + E into its postfix form

Example 5: Use a stack to convert the following infix arithmetic

expression into a postfixexpression. Show the changing status of
the stack in tabular form.
X + Y*Z ^ P (X/Y + Z)

Solution: We would use the algorithm convert( ). The simulation of

this algorithm for the above expression is given in Table 4.8.
The output postfix expression is XYZP^* + XY/Z +
Note: Since a postfix expression has no parentheses, its evaluation
becomes a very easy exercise. In fact, a stack can be suitably used for this
purpose.
4.2.1.4 Evaluation of Postfix Expressions
A stack can be conveniently used for the evaluation
of postfix expressions. The expression is read from left to right. Each
encountered element is examined. If the element is an operand, then the
element is pushed on to the stack. If the element is an operator, then
the two operands are popped from the stack and the desired operation
is done. The result of the operations is again pushed onto the stack.
This process is repeated till the end of the expression (i.e., ;) is
encountered. The final result is popped from the stack.
An algorithm for the evaluation of a postfix expression is given below:
Algorithm evalPostfix( )
{
Step
1.pickanelementfromthepostfixexpression
Table 4.8 Conversion of X + Y*Z ^ P (X/Y + Z)

2.if(elementisanoperand)
if(element!=;)
{pushelementonthestack}
else
{
popresultfromthestack;
printtheresult;
exittostep4;
}
else
{
popoperand1fromstack;
popoperand2fromstack;
performtheoperation;
pushresultonthestack;
}
3.repeatsteps1and2
4.stop

}
Example 6: Write a program that uses a stack to evaluate
a postfix expression.
Solution: The algorithm evalPostfix( )and the following data
structures would be used for the required program.
The postfix expression will be stored in an array called postfix[ ] of
following structure type.

structterm/*structuretostoreanelement*/
{
charelement;
floatval;
};
where
element stores an operand or an operator.
val stores the actual value of the operand or 0 for an operator.
For instance, the expression ABC*+; for values A = 10, B = 15, C = 8 will
be stored in postfix[ ] as shown below:
postfix
An array called stack of type float would also be used.
The required program is given below:
/*Thisprogramusesastacktoevaluateapostfix
expression*/
#include<stdio.h>
#include<conio.h>
structterm/*structuretostoreanelement*/
{
charelement;
floatval;
};
/*arraytostorethepostfixexpression*/
intpush(floatstack[],int*top,floatval,int
size);
floatpop(floatstack[],int*top);
voidmain()

{
floatstack[20];
structtermpostFix[20];
inttop=1;
inti;
intsize;
floatresult;
inttemp;
floatop1,op2;/*theoperands*/
printf(\nEnterthenumberofelementsinthe
postfixexpression);
scanf(%d,&size);
printf(\nEntertheexpressionaselementvalue
pair);
printf(\nEnter0foranoperatorincluding;);
for(i=0;i<size;i++)
{
printf(\nEnterelementvalpair);
fflush(stdin);
scanf(%c%f,&postFix[i].element,
&postFix[i].val);
}
for(i=0;i<size;i++)
{
switch(postFix[i].element)
{
case+:op2=pop(stack,&top);
op1=pop(stack,&top);
result=op1+op2;
temp=push(stack,&top,result,size);
if(temp==0)printf(\nstackfull);
break;
case*:op2=pop(stack,&top);
op1=pop(stack,&top);
result=op1*op2;
temp=push(stack,&top,result,size);
if(temp==0)printf(\nstackfull);
break;
case:op2=pop(stack,&top);
op1=pop(stack,&top);
result=op1op2;

temp=push(stack,&top,result,size);
if(temp==0)printf(\nstackfull);
break;
case/:op2=pop(stack,&top);
op1=pop(stack,&top);
result=op1/op2;
temp=push(stack,&top,result,size);
if(temp==0)printf(\nstackfull);
break;
case;:result=pop(stack,&top);
printf(\nTheresult=%8.2f,result);
break;
default:temp=push(stack,&top,
postFix[i].val,size);
if(temp==0)printf(\nstackfull);
}
}
}
intpush(floatstack[],int*top,floatval,int
size)
{
if(*top<size)
{
*top=*top+1;
stack[*top]=val;
return1;
}
else
return0;
}
floatpop(floatstack[],int*top)
{
floatterm;
term=0;
if(*top>=0)
{
term=stack[*top];
*top=*top1;
returnterm;
}
else

returnterm;
}
For following given input:
A8
B6
C 20
*0
0
;0
the program gives the following output:
The Result = 112.00
Example 7: Use a stack to evaluate the following postfix arithmetic
expression. Show the changing status of the stack in tabular form.
A B C * for A = 8, B = 6, C = 20
Solution: We would use the algorithm evalPostfix( ). The
simulation of this algorithm for the above expression is given in Table
4.9.
Table 4.9 Evaluation of A B C *

Example 8: Use a stack to evaluate the following postfix arithmetic

expression. Show the changing status of the stack in tabular form.
X Y Z P ^ * + A B/C +
=8

for X = 1, Y = 5, Z, = 2, P = 3, A = 15, B = 3, C

Solution: We would use the algorithm evalPostfix( ). The

simulation of this algorithm for the above expression is given in Table
4.10.
Table 4.10 Evaluation of X Y Z P ^ * + A B/C +

4.3 QUEUES

A queue is the most common situation in this world where the first
person in the line is the person to be served first; the new comers join at
the end. We find that in this kind of arrangement, all additions are made
at one end and the deletions made at the other. The end where all

additions are made is called the rear end. The other end from where the
deletions are made is called the front end as shown in Figure 4.8.

Fig. 4.8 The queue

The queue is also a linear data structure of varying size in the sense that
the size of a queue depends upon the number of items currently present
in it. With additions and/ or deletions, the size of the queue increases or
decreases, respectively. Therefore, the queue is also a dynamic data
structure with a capacity to enlarge or shrink.
Consider the queue of processes shown in Figure 4.9. The operating
system maintains a queue for scheduled processes that are ready to run.
The dispatcher picks the process from the front end of the queue and
dispatches it to CPU and a new process joins at the rear end of the queue.

Fig. 4.9 Queue of processes maintained in the main memory

It may be noted that when the new processes join in, the size of the
queue increases whereas the size of the queue will reduce when
processes finish their job and leave the system.
4.3.1 Queue Operations
A queue, being a linear data structure, can be comfortably mapped on a
one-dimensional array. Let us consider a queue maintained in an array
called Queue[N], shown in Figure 4.10. A variable calledFront is
pointing to the logical front end of the queue from where the removal or
deletion of an element takes place. Currently, Front is pointing to an
empty location in the queue. Similarly, a variable called Rear is pointing
to the logical rear end of the queue where the new elements can be
added. Currently, Rear is pointing to a location containing the item G.

Fig. 4.10 An array representation of a queue

Let us now add an item H into the queue. Before the item is added into
the queue, the Rear is incremented by one to point to the next vacant
location as shown in Figure 4.11. The new item H is added at location
currently being pointed by Rear.

Fig. 4.11 Addition of the item H at the rear end

It may be noted that if more additions are made on the queue, then a
stage would come when Rear = N 1, i.e., last location of the array. Now,
no more additions can be performed on the queue. This condition (Rear
= N 1) is called QueueFull.
Let us now delete an item from the queue. Before an item is deleted or
removed, the Front is incremented by one to point to the next location,
i.e., the location containing item A. The item (i.e., A) is then deleted as
shown in Figure 4.12.

Fig. 4.12 Deletion of the item A from the front end

It may be noted that if more deletions are done on this queue, then a
stage would reach when there will be no items on the queue,
i.e., Front = Rear as shown in Figure 4.13. This condition
(Front = Rear) is called Queue_empty.

Fig. 4.13 No item in the queue

The algorithms for addition and deletion operations on the queue are
given below. These algorithms use an array called Queue[N] to represent
a queue of size N locations. A variable called Front keeps track of the
front of the queue, i.e., the location from where deletions are made.
Another variable called Rear keeps track of the rear end of the queue,
i.e., the location where additions are made. Another variable called item
is used to store the item to be added or deleted from the queue.
The algorithm for addition is given below:
Algorithm addQ( )
{
Step
1.if(Rear>=N)then{prompt(QueueFull);
exit}
2.Rear=Rear+1;
3.Queue[Rear]=item;
4.stop
}
It may be noted that the item is added to the queue (i.e., addition
operation) provided the queue is not already full, i.e., it is checked at
step1 whether or not the queue has room for another item.
The algorithm for deletion is given below:
Algorithm delQ( )
{
Step
1.if(Front==Rear)thenprompt(QueueEmpty);
else
Front=Front+1
Item=Queue[Front];
2.stop
}
Note that in the above algorithm, before removing an item from the
queue, it is checked at step 1 to see if there is at least one item on the
queue that can be removed.
The performance of the queue data structure for n elements is given
below:
o The space complexity is O(n).
o The time complexity of each operation is O(1).
o The size of the queue, implemented using array, must be defined in
advance, i.e., a priori. Moreover, the size cannot be changed.

Example 9: Write a program in C that uses the following menu to

simulate the queue operations on items of inttype.

Enter your choice:

Solution: The above defined algorithms addQ( ) and delQ( ) would be
employed to write the program. The menu would be created
using printf( ) statements and displayed within a do-while loop. In
fact, a do-while loop is the most appropriate loop for display and
manipulations of menu items.
Note:
o The addQ( ) function would return 1 if the operation is successful and 0
otherwise, i.e, 0 indicates that the queue was full.
o The delQ( ) function would return the deleted value if the operation is
successful and 9999 otherwise, i.e., the unusual value 9999 indicates that
the queue was empty.

The required program is given below:

/*Thisprogramsimulatesqueueoperations*/
#include<stdio.h>
#include<conio.h>
intaddQ(intQueue[],int*Rear,intval,intsize);
intdelQ(intQueue[],int*Front,int*Rear);
voiddisplay(intQueue[],intFront,intRear);
voidmain()
{
intQueue[20];/*thequeuedeclaration*/
intFront;
intRear;
intval;
intsize;
intchoice;
intresult;
Front=Rear=0;/*initiallythequeuesettobe
empty*/

printf(\nEnterthesizeofthequeue);
scanf(%d,&size);
size;/*adjustedforindex=0*/
/*createmenu*/
do
{
clrscr();
printf(\nMenuQueueOperations);
printf(\nAdd1);
printf(\ndelete2);
printf(\nDisplayqueue3);
printf(\nQuit4);
printf(\nEnteryourchoice:);
scanf(%d,&choice);
switch(choice)
{
case1:printf(\nEnterthevaluetobe
added);
scanf(%d,&val);
result=addQ(Queue,&Rear,val,size);
if(result==0)
printf(\nThequeuefull);
break;
case2:result=delQ(Queue,&Front,&Rear);
if(result==9999)
printf(\nThequeueisEmpty);
else
printf(\nThedeletedvalue=%d,result);
break;
case3:display(Queue,Front,Rear);
break;
}
printf(\n\nPressanykeytocontinue);
getch();
}
while(choice!=4);
}
intaddQ(intQueue[],int*Rear,intval,intsize)
{
if(*Rear>=size)
return0;/*thequeueisFull*/

else
{
*Rear=*Rear+1;
Queue[*Rear]=val;
return1;
}
}
intdelQ(intQueue[],int*Front,int*Rear)
{intval;
if(*Front==*Rear)
return9999;/*thequeueisempty*/
else
{
*Front=*Front+1;
val=Queue[*Front];
returnval;
}
}
voiddisplay(intQueue[],intFront,intRear)
{
inti;
printf(\nThecontentsofqueueare:);
for(i=Front+1;i<=Rear;i++)
printf(%d,Queue[i]);
}
Note:
1. The drawback of a linear queue is that a stage may arrive when both Front
and Rear are equal and point to the last location of the array as shown
in Figure 4.14.

Fig. 4.14 The queue is empty as well as full

Now, the queue is useless because the queue is full (Rear > = N 1) as
well as empty (Front =Rear). Thus, no additions and deletions can be
performed. This situation can be handled by resetting Front and Rear to 0th
location, i.e., Front = Rear = 0.

2. The major drawback of linear queue is that at a given time, all the locations
to the left of Front are always vacant and unutilized.

4.3.2 Circular Queue

The drawback of the linear queues, as discussed above, can be removed
by making the queue spill over the elements from last position of array to
the 0th element. This arrangement of elements makes the queue
as circularqueue wherein as soon as Rear or Front exceeds N 1, it
is set to 0. Now, there is no end of the queue. The Rear and Front will
move around in the queue in the circle and no locations shall be left
vacant.
Let us add an item C to the queue of Figure 4.14. The Rear is pointing to
location N1. After incrementing Rear (i.e., Rear = Rear + 1 = N 1 + 1
= N), we find that it is pointing to a location equal to N, which does not
exist. Therefore, Rear is set to location 0, the beginning of the array and
the item C is stored at this location pointed by Rear as shown in Figure
4.15.
Now, this arrangement ensures that there will be no space left unutilized
in the queue. For instance, let us add one by one the items I,R,
C,U,L,A and R to the queue (see Figure 4.16). We find
that the locations left of Front are now being filled, removing the
drawback of linear queues.

Fig. 4.15 The rear is made to move to the 0th location.

It may be noted that after more additions, a stage will come when Rear
becomes equal to Front (Rear= Front), indicating that the queue is full.

Fig. 4.16 More additions to circular queue

Let us now delete an element from the queue. The Front is pointing to
location N1. After incrementing Front (i.e., Front=Front+1=
N1+1=N), we find that it is pointing to a location equal to N,
which does not exist. Therefore, Front is set to location 0, the beginning

of the array and the item C is deleted from this location pointed by
Front as shown in Figure 4.17.
It may be noted that after more deletions, a stage will come when Front
becomes equal to Rear (Front= Rear), indicating that the queue is
empty.

Fig. 4.17 For deletion, Front is set to 0th location

It may be further noted that a very interesting situation has turned up,
i.e., for both eventsqueue full and emptythe condition to be
tested is (Rear = Front). This conflict can be resolved by the following
decisions:
o The queue empty condition remains the same, i.e., Front == Rear.
o In case of queue full, the condition Rear+1==Front would be
tested.
o In this arrangement one location, being pointed by Front, shall remain
vacant.
o The following formulae would be used for automatic movement of the
variables Front; andRear within the circular queue:

Rear=(Rear+1)modN
Front=(Front+1)modN
where N is the size of the queue.
The algorithms for addition and deletion operations in a circular queue
are given below. These algorithms use an array called cirQ[N] to
represent a queue of size N locations. A variable called Front keeps track
of the front of the queue, i.e., the location from where deletions are
made. Another variable called Rear keeps track of the rear end of the
queue, i.e., the location where additions are made. Another variable
called item is used to store the item to be added or deleted from the
queue.
The algorithm for addition is given below:
Algorithm addQ()
{
Step
1.if(((Rear+1)%N)==Front)

then{prompt(QueueFull);exit}
2.Rear=(Rear+1)%N;
3.cirQ[Rear]=item;
4.stop
}
It may be noted that the item is added to the queue (i.e., addition
operation) provided the queue is not already full, i.e., at step1 it is
checked whether or not the queue has room for another item.
The algorithm for deletion is given below:
Algorithm delQ()
{
Step
1.if(Front==Rear)thenprompt(QueueEmpty);
else
Front=(Front+1)%N
Item=Queue[Front];
returnItem;
2.stop
}
Note that in the above algorithm, before removing an item from the
queue, it is checked at step 1 to see if there is at least one item on the
queue that can be removed.
Example 10: Write a program in C that uses the following menu to
simulate the circular queue operations on items of int type.

Enter your choice:

Solution: The above defined algorithms addQ() and delQ() for
circular queue would be employed to write the program. The menu
would be created using printf() statements and displayed within a dowhile loop.
Note:

o The addQ() function would return 1 if the operation is successful and 0

otherwise, i.e., 0 indicates that the queue was full.
o The delQ() function would return the deleted value if the operation is
successful and 9999 otherwise, i.e., the unusual value 9999 indicates that
the queue was empty.

The required program is given below:

/*ThisprogramsimulatesCircularQueueoperations
*/
#include<stdio.h>
#include<conio.h>
intaddQ(intcirQ[],int*Front,int*Rear,intval,
intsize);
intdelQ(intcirQ[],int*Front,int*Rear,int
size);
voiddisplay(intcirQ[],intFront,intRear,int
size);
voidmain()
{
intcirQ[20];/*thequeuedeclaration*/
intFront;
intRear;
intval;
intsize;
intchoice;
intresult;
Front=Rear=0;/*initiallythequeuesettobe
empty*/
printf(\nEnterthesizeofthequeue);
scanf(%d,&size);
/*createmenu*/
do
{
clrscr();
printf(\nMenuCircularqueueoperations);
printf(\nAdd1);
printf(\nDelete2);
printf(\nDisplayqueue3);
printf(\nQuit4);
printf(\nEnteryourchoice:);
scanf(%d,&choice);

switch(choice)
{
case1:printf(\nEnterthevaluetobe
added);
scanf(%d,&val);
result=addQ(cirQ,&Front,&Rear,val,size);
if(result==0)
printf(\nThequeuefull);
break;
case2:result=delQ(cirQ,&Front,&Rear,size);
if(result==9999)
printf(\nThequeueisEmpty);
else
printf(\nThedeletedvalue=%d,result);
break;
case3:display(cirQ,Front,Rear,size);
break;
}
printf(\n\nPressanykeytocontinue);
getch();
}
while(choice!=4);
}
intaddQ(intcirQ[],int*Front,int*Rear,intval,
intsize)
{
if(((*Rear+1)%size)==*Front)
return0;/*thequeueisFull*/
else
{
*Rear=(*Rear+1)%size;
cirQ[*Rear]=val;
return1;
}
}
intdelQ(intcirQ[],int*Front,int*Rear,int
size)
{intval;
if(*Front==*Rear)
return9999;/*thequeueisempty*/
else

{
*Front=(*Front+1)%size;
val=cirQ[*Front];
returnval;
}
}
voiddisplay(intcirQ[],intFront,intRear,int
size)
{
inti;
printf(\nThecontentsofqueueare:);
i=Front;
while(i!=Rear)
{
i=(i+1)%size;
printf(%d,cirQ[i]);
}
}
It may be noted that the circular queue has been implemented in a onedimensional array, which is linear by nature. We only view the array to
be a circle in the sense that it is imagined that the last location of the
array is immediately followed by the 0th location as shown in Figure
4.18.
The shaded area shows the queue elements. When the queue is full, then
it satisfies the condition: (Rear + 1) % N == Front. The queue empty
condition is indicated by the condition: Front == Rear. However, at
least one location remains vacant all the time.

Fig. 4.18 The circular queue

4.3.3 Priority Queue

Sometimes, a queue does not operate on strictly first in first out (FIFO)
basis. On the contrary, it stores the elements according to the priority
associated with each stored element. In a post office, the airmail has a
higher priority than the speed post and speed post in turn has more
priority than an ordinary mail. Therefore, the post office makes a queue
of mail ordered by their priority, i.e., all airmail letters are placed at the
head of the queue followed by speed post letters. The speed post letters
are followed by registered parcels and so on. The ordinary mail with least
priority lies at the end of the queue.
Let us take an example from operating system. A priority scheduling
policy would store the processes as per their priority in a queue. The
scheduler will add a new process to the queue according to its priority
and the highest priority process would be served first as shown in Figure
4.19.
In fact, the operating system will have to maintain such priority queues
at all non-shareable devices such as printers, plotters, etc.
Note: Since every element in the queue is ordered on priority, an entry
into the priority queue has to be a pair (e, p) where e is the name of the
element and p is its priority. In case of records of items, the e is chosen
to be a field or key of the record.

Fig. 4.19 The priority queue maintained by operating system

Consider the job pool given in Table 4. 11. The processes opened by the
operating system have been assigned priorities.
Table 4.11 Job pool
process no.

priority

P8

P5

P7

P4

P2

Now, the scheduler shall create a priority queue that can store a list of
pairs (Process No., priority). It has two possible choices to store the pairs
as shown in Figure 4.20.
Choice (a): The addition operation simply adds the newly arrived job at
the rear end, i.e., in the order of its arrival. This operation is O(1).
The deletion operation requires the process with higher priority to be
searched and removed, P7 in this case. This operation is O(n) for n
elements in the queue as we have to traverse the queue to find the
element.

Fig. 4.20 The choices to store the job pool

Choice (b): The addition operation requires the process with higher
priority to be inserted at its proper position in the queue. This operation
is O(n) for n elements in the queue as the location has to be found where
the insertion can take place.
The deletion operation simply removes the process from the front of the
queue because the process with highest priority is already at the front of
the queue. This operation is O(1).
Let us now use the choice (b) to implement a priority queue for processes
called proc of following structure type
structproc
{
charprocess[3];
intpriority;
};
The following menu would be used to interact with the users:

Enter your choice:

A function called addPq() inserts an ith process pi with
priority pr at its proper position in the priority queue represented
using the array pQ[N] where N is the size of the queue. A
variableRear points to rear end of the priority queue, i.e., the least
priority process.
A function called delPq() removes a process from the
current Front of the priority queue.
The required program is given below:
/*Thisprogramsimulatespriorityqueueoperations
*/
#include<stdio.h>
#include<conio.h>
structproc{
charprocess[3];
intpriority;
};
intaddPq(structprocpQ[],int*Rear,structproc
val,intsize);
intdelPq(structprocpQ[],int*Front,int*Rear,
structproc*val);
voiddisplay(structprocpQ[],intFront,intRear);
voidmain()
{
structprocpQ[20];/*thepriorityqueue
declaration*/
intFront;
intRear;
structprocval;
intsize;
intchoice;
intresult;

Front=Rear=0;/*initiallythequeuesettobe
empty*/
printf(\nEnterthesizeofthequeue);
scanf(%d,&size);
/*createmenu*/
do
{
clrscr();
printf(\nMenuPriorityqueueoperations);
printf(\nAdd1);
printf(\nDelete2);
printf(\nDisplayqueue3);
printf(\nQuit4);
printf(\nEnteryourchoice:);
scanf(%d,&choice);
switch(choice)
{
case1:printf(\nEntertheprocessnumber
anditsprioritytobeadded);
scanf(%s%d,&val.process,
&val.priority);
result=addPq(pQ,&Rear,val,size);
if(result==0)
{printf(\nThequeuefull);}
break;
case2:result=delPq(pQ,&Front,&Rear,
&val);
if(result==0)
printf(\nThequeueisempty);
else
printf(\nThedeletedproc=%swith
priority%d,val.
process,val.priority);
break;
case3:display(pQ,Front,Rear);
break;
}
printf(\n\nPressanykeytocontinue);
getch();
}
while(choice!=4);

}
intaddPq(structprocpQ[],int*Rear,structproc
val,intsize)
{
inti;
if((*Rear+1)>=size)
return0;/*thequeueisFull*/
else
{
*Rear=(*Rear+1);
i=*Rear1;
while(val.priority>pQ[i].priority)
{
pQ[i+1]=pQ[i];
i=i1;
}
i++;
pQ[i]=val;
}
return1;
}
intdelPq(structprocpQ[],int*Front,int*Rear,
structproc*val)
{
if(*Front==*Rear)
return0;/*thequeueisempty*/
else
{
*Front=(*Front+1);
*val=pQ[*Front];
return1;
}
}
voiddisplay(structprocpQ[],intFront,intRear)
{
inti;
printf(\nThecontentsofqueueare:);
i=Front;
while(i!=Rear)
{
i=i+1;

printf((%s,%d),pQ[i].process,
pQ[i].priority);
}
}
It may be noted that in order to keep it simple, the priority queue has
been implemented as a linear queue. A sample input is given below:
Enter the process number and its priority to be added P4 7
where P4 is the name of the process and 7 is its priority.
4.3.4 The Deque
A deque is a queue that allows additions and deletions from both ends
(see Figure 4.21). This data structure is suitable for implementing the
redo and undo operations of software such as word processor, paint
program, drawing tools, etc.
A one-dimensional array can be suitably used to implement a deque
wherein the following four operations need to be designed:
1.
2.
3.
4.

addRear(): adds an item at the rear end

delRear(): removes an item from the rear end
addFront(): adds an item at the front end
delFront(): removes an item from the front end

Fig. 4.21 The deque

Some programmers view a deque as two stacks connected back to back
as shown in Figure 4.22.
For deque arrangement shown in Figure 4.22, the following four
operations need to be designed:
1.
2.
3.
4.

pushRear(): adds an item at the rear end

popRear(): removes an item from the rear end
pushFront(): adds an item at the front end
popFront(): removes an item from the front end

Fig. 4.22 The deque (stack view)

Example 11: Write a program in C that uses the following menu to
simulate the deque operations on items of int type.

Enter your choice:

Solution: A one-dimensional array called dQ[] would be used to
implement the deque wherein the following four functions would be
used. The algorithms for these functions are same as used in a linear
queue.
1.
2.
3.
4.

addRear(): adds an item at the rear end

delRear(): removes an item from the rear end
addFront(): adds an item at the front end
delFront(): removes an item from the front end

The menu would be created using printf() statements and displayed

within a dowhile loop.
The required program is given below:
/*ThisprogramsimulatesdeQueoperations*/
#include<stdio.h>
#include<conio.h>
intaddRear(intdQ[],int*Rear,intval,intsize);
intaddFront(intdQ[],int*Front,intval);
intdelRear(intdQ[],int*Rear,int*Front,int
*val);

intdelFront(intdQ[],int*Front,int*Rear,int
*val);
voiddisplay(intdQ[],intFront,intRear);
voidmain()
{
intdQ[20];/*thequeuedeclaration*/
intFront;
intRear;
intval;
intsize;
intchoice;
intresult;
Front=Rear=0;/*initiallythequeuesettobe
empty*/
printf(\nEnterthesizeofthequeue);
scanf(%d,&size);
size;/*adjustedforindex=0*/
/*createmenu*/
do
{
clrscr();
printf(\nMenudQueOperations);
printf(\nAdditematRear1);
printf(\nAdditematFront2);
printf(\nDeleteitemfromRear3);
printf(\nDeleteitemfromFront4);
printf(\nDisplaydeque5);
printf(\nQuit6);
printf(\nEnteryourchoice:);
scanf(%d,&choice);
switch(choice)
{
case1:printf(\nEnterthevaluetobeadded);
scanf(%d,&val);
result=addRear(dQ,&Rear,val,size);
if(result==0)
printf(\nThequeueisfull);
break;
case2:printf(\nEnterthevaluetobeadded);
scanf(%d,&val);
result=addFront(dQ,&Front,val);

if(result==0)
printf(\nThequeueisfull);
break;
case3:result=delRear(dQ,&Rear,&Front,&val);
if(result==0)
printf(\nThequeueisempty);
else
printf(\nThedeletedvalue=%d,val);
break;
case4:result=delFront(dQ,&Front,&Rear,&val);
if(result==0)
printf(\nThequeueisempty);
else
printf(\nThedeletedvalue=%d,val);
break;
case5:display(dQ,Front,Rear);
break;
}
printf(\n\nPressanykeytocontinue);
getch();
}
while(choice!=6);
}
intaddRear(intdQ[],int*Rear,intval,intsize)
{
if(*Rear>=size)
return0;/*thequeueisFull*/
else
{
*Rear=*Rear+1;
dQ[*Rear]=val;
return1;
}
}
intaddFront(intdQ[],int*Front,intval)
{
if(*Front<=0)
return0;/*thequeueisfull*/
else
{
dQ[*Front]=val;

*Front=*Front1;
return1;
}
}
intdelRear(intdQ[],int*Rear,int*Front,int
*val)
{
if(*Rear==*Front)
return0;/*thequeueisempty*/
else
{
*val=dQ[*Rear];
*Rear=*Rear1;
return1;
}
}
intdelFront(intdQ[],int*Front,int*Rear,int
*val)
{
if(*Front==*Rear)
return0;/*thequeueisempty*/
else
{
*Front=*Front+1;
*val=dQ[*Front];
return1;
}
}
voiddisplay(intQueue[],intFront,intRear)
{
inti;
printf(\nThecontentsofqueueare:);
for(i=Front+1;i<=Rear;i++)
printf(%d,Queue[i]);
}
Example 12: The items can be iteratively removed from both ends of a
deque and compared with each other. Therefore, the deque can be easily
used to determine whether a given string is a palindrome or not. Write
a program in C that uses a deque to determine whether a given string is
a palindrome or not.

Solution: An array called String[N] would be used to store the string.

As one location in a queue necessarily remains vacant, an array
called dQ[N+1] would be used to implement a deque. The functions
developed in Example11 would be used to do the required task, i.e., to
determine whether the input string is palindrome or not.
The required program is given below:
/*ThisprogramusesdeQuetodeterminewhethera
givenstringisapalindromeornot*/
#include<stdio.h>
#include<conio.h>
intaddRear(chardQ[],int*Rear,charval,int
size);
intdelRear(chardQ[],int*Rear,int*Front,char
*val);
intdelFront(chardQ[],int*Front,int*Rear,char
*val);
intaddFront(chardQ[],int*Front,charval);
voiddisplay(chardQ[],intFront,intRear);
voidmain()
{
charString[10];
chardQ[11];/*thequeuedeclaration*/
intFront;
intRear;
charval,val1,val2;
inti,size;
intflag;
intresult;
printf(\nEnterthestring);
scanf(%s,String);
i=0;
size=10;/*adjustedforindex=0*/
Front=Rear=0;/*initiallydequeisempty*/
while(String[i]!=\0)/*Addthestringto
deque*/
{
result=addRear(dQ,&Rear,String[i],size);
if(result==0)

printf(\nThequeuefull);
i++;
}
size=i1;
flag=1;/*Assumingthatthestringis
while(flag&&result)
{
result=delRear(dQ,&Rear,&Front,&val1);
result=delFront(dQ,&Front,&Rear,&val2);
if(val1!=val2){flag=0;break;}/*stringis
notapalindrome*/
}
if(flag==1)printf(\nThestringispalindrome);
elseprintf(\nThestringisnotapalindrome);
}
intaddRear(chardQ[],int*Rear,charval,intsize)
{
if(*Rear>=size)
return0;/*thequeueisfull*/
else
{
*Rear=*Rear+1;
dQ[*Rear]=val;
return1;
}
}
intaddFront(chardQ[],int*Front,charval)
{
if(*Front<=0)
return0;/*thequeueisfull*/
else
{
dQ[*Front]=val;
*Front=*Front1;
return1;
}
}
intdelRear(chardQ[],int*Rear,int*Front,char
*val)
{
if(*Rear==*Front)

return0;/*thequeueisempty*/
else
{
*val=dQ[*Rear];
/*Ifthesizeofstringisoddthendonot
decrementRear*/
if((*Rear1)>*Front)*Rear=*Rear1;
return1;
}
}
intdelFront(chardQ[],int*Front,int*Rear,char
*val)
{
if(*Front==*Rear)
{
return0;
}/*thequeueisempty*/
else
{
*Front=*Front+1;
*val=dQ[*Front];
return1;
}
}
voiddisplay(charQueue[],intFront,intRear)
{
inti;
printf(\nThecontentsofqueueare:);
for(i=Front+1;i<=Rear;i++)
printf(%c,Queue[i]);
}
Note: The function addFront() of deque system is redundant in this
application. The functiondelRear() has been suitably modified for
strings having odd number of characters.
Example 13: Two stacks can be implemented using a single array.
Write a program in C that allows push and pop operations from the
specified stack. It must also check the conditions stackemptyand
full for a specified stack.

Solution: In a single array, two stacks can be implemented by making

them grow from both ends towards the middle as shown in Figure 4.23.
It may be noted that 0th element becomes the bottommost element of 1st
stack, i.e., stack1. Similarly, the (N 1)th element becomes the
bottommost element of the 2nd stack, i.e., stack2.
An array called twoStack[N] would be used to implement the required
task. Two variables called Top1and Top2 would keep track of the top
locations of stack1 and stack2, respectively.
Initially, Top1 andTop2 would be set to locations 1 and N, respectively
to suggest that both the stacks are initially empty.
Note: The stack full condition for both the stacks would be: *top1 + 1
== *top2

Fig. 4.23 Two stacks that grow towards middle of the array
The required program is given below:
/*Thisprogramsimulatestwostacksoperations*/
#include<stdio.h>
#include<conio.h>
intpushStack1(intstak[],int*top1,int*top2,int
val);
intpushStack2(intstak[],int*top1,int*top2,int
val);
intpopStack1(intstak[],int*top1);
intpopStack2(intstak[],int*top2);
voiddispStack1(intstack[],inttop1);
voiddispStack2(intstack[],inttop2);
voidmain()
{
inttwoStack[10];/*thestackdeclaration*/
inttop1=1;/*initiallythestack1isempty*/
inttop2=10;/*initiallythestack2isempty*/
intval;
intchoice;
intresult;
/*createmenu*/
do

{
clrscr();
printf(\nMenutwoStackOperations);
printf(\nPushStack11);
printf(\nPushStack22);
printf(\nPopStack13);
printf(\nPopStack24);
printf(\nDispStack15);
printf(\nDispStack26);
printf(\nQuit7);
printf(\nEnteryourchoice:);
scanf(%d,&choice);
switch(choice)
{
case1:printf(\nEnterthevaluetobe
pushed);
scanf(%d,&val);
result=pushStack1
(twoStack,&top1,&top2,val);
if(result==0)
printf(\nThestackisfull);
break;
case2:printf(\nEnterthevaluetobe
pushed);
scanf(%d,&val);
result=pushStack2(twoStack,&top1,&top2,
val);
if(result==0)
printf(\nThestackisfull);
break;
case3:result=popStack1(twoStack,&top1);
if(result==9999)
printf(\nThestackisempty);
else
printf(\nThepoppedvalue=%d,result);
break;
case4:result=popStack2(twoStack,&top2);
if(result==9999)
printf(\nThestackisempty);
else
printf(\nThepoppedvalue=%d,result);

break;
case5:dispStack1(twoStack,top1);
break;
case6:dispStack2(twoStack,top2);
break;
}
printf(\n\nPressanykeytocontinue);
getch();
}
while(choice!=7);
}
intpushStack1(intstack[],int*top1,int*top2,
intval)
{
if(*top1+1==*top2)
return0;/*thestackisfull*/
else
{
*top1=*top1+1;
stack[*top1]=val;
return1;
}
}
intpushStack2(intstack[],int*top1,int*top2,int
val)
{
if(*top1+1==*top2)
return0;/*thestackisfull*/
else
{
*top2=*top21;
stack[*top2]=val;
return1;
}
}
intpopStack1(intstack[],int*top1)
{intval;
if(*top1<0)
return9999;/*thestackisempty*/
else
{

val=stack[*top1];
*top1=*top11;
returnval;
}
}
intpopStack2(intstack[],int*top2)
{intval;
if(*top2>10)
return9999;/*thestackisempty*/
else
{
val=stack[*top2];
*top2=*top2+1;
returnval;
}
}
voiddispStack1(intstack[],inttop1)
{
inti;
printf(\nThecontentsofstack1are:);
for(i=top1;i>=0;i)
printf(%d,stack[i]);
}
voiddispStack2(intstack[],inttop2)
{
inti;
printf(\nThecontentsofstack2are:);
for(i=top2;i<=10;i++)
printf(%d,stack[i]);
}
EXERCISES
1. Give static implementation of stack by writing push and pop routine for it.
2. Explain overflow and underflow conditions of a stack with examples.
3. How can a stack be used in checking the well-formedness of an
expression, i.e., balance of the left and right parenthesis of the expression?
4. Write down an algorithm to implement two stacks using only one array.
Your stack routine should not report an overflow unless every slot in the
array is used.
5. Write a non-recursive program to reverse a string using stack.
6. Explain prefix, infix, and postfix expressions with examples.

7. Describe a method to convert an infix expression in to

a postfix expression with the help of a suitable example.
8. Translate following infix expressions to their
equivalent postfix expressions:

1. (x + y z)/(h + k)*s
2. j - k/g^h + (n + m)
3. a*(b c)/d + e*f
2 Write a program that evaluates an expression represented
in postfix form.
3 Evaluate the following postfix expressions for the provided data:

1. ab ^ c * d / e + where a = 5, b = 3, c = d = 2, e = 9
2. abcde + * + where a 5 + , b = 4, c = 7, d = 5, e = 2
3. ab + cd * + e * where a = 2, b = 6, c = 3, d = 5, e = 9
2 What are the applications of a stack? Support your answer with suitable
examples.
3 What is a linear queue? What are the limitations of linear queue?
4 What do you understand by a queue? Write algorithm for inserting and
deleting of a data element from the queue.
5 Write the functions insert() and delete() for a circular queue.
6 Write a menu driven main() program that tests the functions developed in
Exercise 14.
7 Differentiate between linear queue and circular queue. Which one is better
and why?
8 Describe the applications of queues.
9 What are priority queues? Write their applications.
10 What is doubly ended queue? Write a program/algorithm to implement a
doubly ended queue (deque).

5
Pointers
CHAPTER OUTLINE

5.1 Introduction
5.2 Pointer Variables
5.3 Pointers and Arrays
5.4 Array of Pointers
5.5 Pointers and Structures
5.6 Dynamic Allocation
5.1 INTRODUCTION

Pointers are the most powerful feature of C. The beginners of C find

pointers hard to understand and manipulate. In an attempt to unveil the
mystery behind this aspect, the basics of pointers which generally remain
in background are being discussed in the following sections.
5.1.1 The & Operator
When a variable x is declared in a program, a storage location in the
main memory is made available by the compiler. For example, Figure
5.1 shows the declaration of x as an integer variable and its storage
location in the main memory.

Fig. 5.1 The variable x and its equivalent representation in memory

From Figure 5.1, it may be observed that x is the name associated by the
compiler to a location in the memory of the computer. Let us assume
that, at the time of execution, the physical address of this memory
location (called x)is 2712. Now, a point worth noting is that this
memory location is viewed by the programmer as variable x and by the
operating system as an address 2712. The address of variable x can be
obtained by & an address of operator. This operator when applied to a
variable gives the physical memory address of the variable. Thus, &x will
provide the address 2712. Consider the following program segment:

Fig. 5.2 The contents of variable x

x=15;
printf(\nvalueofx=%d,x);
printf(\naddressofx=%d,&x);
The output of the above program segment would be as shown below:
Valueofx=15
Addressofx=2712(assumedvalue)

The contents of variable x (memory location 2712) are shown in Figure

5.2.
It may be noted here that the value of address of x (i.e., 2712) is assumed
and the actual value is dependent on machine and execution time.
5.1.2 The * Operator
The * is an indirection operator or valueataddress
operator. In simple words, we can say that if address of a variable is
known, then the * operator provides the contents of the variable.
Consider the following program segment:
#include<stdio.h>
main()
{
intx=15;
printf(\nValueofx=%d,x);
printf(\nAddressofx=%u,&x);
printf(\nValueataddress%d=%d,*(&x));
}
The output of the above program segment would be as shown below:
Valueofx=15
Addressofx=2712
Valueataddress2712=15
5.2 POINTER VARIABLES

An address of a variable can be stored in a special variable called pointer

variable. A pointer variable contains the address of another variable. Let
us assume that y is such a variable. Now, the address of variable x can be
assigned to y by the statement: y=&x. The effect of this statement is
shown inFigure 5.3.

Fig. 5.3 The address of x is being stored in pointer variable y

From Figure 5.3, it is clear that y is the variable that contains the address
(i.e., 2712) of another variable x, whereas the address of y itself is 1232
(assumed value). In other words, we can say that y is a pointer to
variable x. See Figure 5.4.

Fig. 5.4 y is a pointer to x

A pointer variable y can be declared in C as shown below:
int*y;
The above declaration means that y is a pointer to a variable of
type int. Similarly, consider the following declaration:

The above declaration means that p is a pointer to a variable of

type char.
Consider the following program segment:
#include<stdio.h>
main()
{
intx=15;
int*y;
y=&x;
printf(\nValueofx=%d,x);
printf(\nAddressofx=%u,&x);
printf(\nValueofx=%d,*y);
printf(\nAddressofx=%u,y);
printf(\nAddressofy=%u,&y);
}
The output of the above program segment would be as shown below:
Valueofx=15
Addressofx=2712
Valueofx=15
Addressofx=2712
Addressofy=1232
Let us, now, consider the following statements:
floatval=35.67;
float*pt;

The first statement declares val to be a variable of type float and

initializes it by value 35.67. The second statement declares pt to be
pointer to a variable of type float. Please notice that pt can point to a
variable of type float but it is currently not pointing to any variable as
shown in Figure 5.5.

Fig. 5.5 The pointer pt is not pointing to any variable

The pointer pt can point to the variable val by assigning the address
of val to pt as shown below:
pt=&val;
Once the above statement is executed, the pointer pt gets the address
of val and we can say that logically, pt points to the variable val as
shown in Figure 5.6.

Fig. 5.6 The pointer pt pointing to variable val

However, one must keep in mind that the arrow is for
illustration sake. Actually, the pointer pt contains the address
of val. The address is an integer which is not a simple value but a
reference to a location in the memory.
Examples of some valid pointer declarations are:
(i)int*p;
(ii)char*i,*k;
(iii)float*pt;
(iv)int**ptr;
The example (iv) means the ptr is a pointer to a location which itself is a
pointer to a variable of type integer.
It may be noted here that a pointer can also be incremented to point to
an immediately next location of its type. For example, if the contents of a
pointer p of type integer are 5224 then the content of p++ will be 5226
instead of 5225 (see Figure 5.7). The reason being that an int is always of
2 bytes size and, therefore, stored in two memory locations as shown
in Figure 5.7

Fig. 5.7 The variable of type int occupies two bytes

The amount of storage taken by various types of data is tabulated
in Table 5.1.
Table 5.1 Amount of storage taken by data types
Data type

Amount of storage

character

1 byte

integer

2 byte

float

4 byte

Rong

4 byte

double

8 byte

Thus, a pointer of float type will point to an address of 4 bytes of location

and, therefore, an increment to this pointer will increment its contents
by 4 locations. It may be further noted that more than one pointer can
point to the same location. Consider the following program segment:
intx=37
int*p1,*p2;
The following statement:
p1=&x;
Makes the pointer p1 point to variable x as shown in Figure 5.8.
Fig. 5.8 Pointer p1 points to variable x
The following statement:

p2=p1
Makes p2 point to the same location which is being pointed by p1 as
shown in Figure 5.9.

Fig. 5.9 Pointers p1 and p2 pointing to x

The contents of variable x are now reachable by both the pointers p1
andp2. Thus, two or more entities can cooperatively share a single
memory structure with the help of pointers. In fact, this technique can be
used to provide efficient communication between different parts of a
program.
The pointers can be used for a variety of purposes in a program; some of
them are as follows:
o To pass address of variables from one function to another.
o To return more than one value from a function to the calling function.
o For the creation of linked structures such as linked lists and trees.

Example 1: Write a program that reads from the keyboard two

variables xandy of type integer. It prints the exchanged contents of
these variables with the help of pointers without altering the variables.
Solution: We will use two pointers p1andp2 for variables xand
y. A third pointer p3 will be used as a temporary pointer for the
exchange of pointers p1andp2. The scheme is shown in Figure 5.10.
The program is given below:
/*Thisprogramillustratestheusageofpointersto
exchangethecontentsoftwovariables*/
#include<stdio.h>
main()
{
intx,y;
int*p1,*p2,*p3;/*pointerstointegers*/
printf(\nEntertwointegervalues);
scanf(%d%d,&x,&y);
/*Assigntheaddressesxandytop1andp2*/

p1=&x;
p2=&y;
/*Exchangethepointers*/
p3=p1;
p1=p2;
p2=p3;
/*Printthecontentsthroughexchangedcontents*/
printf(\nTheexchangedcontentsare);
printf(%d&%d,*p1,*p2);
}

Fig. 5.10 Exchange of contents of variables by exchanging pointers

Note: The output would show the exchanged values because of exchange
of pointers (see Figure 5.10) whereas the contents of xandy have
remained unaltered.
5.2.1 Dangling pointers
A dangling pointer is a pointer which has been allocated but does not
point to any entity. Such an un-initialized pointer is dangerous in the
sense that a dereference operation on it will result in an unpredictable
operation or, may be, a runtime error.
int*ptr;
*ptr=50;
The first statement allocates the pointer called ptr as shown in Figure
5.11. It may be noted that, currently, ptr is a dangling pointer. The
second statement is trying to load a value (i.e., 50) to a location which is
pointed by ptr. Obviously, this is a dangerous situation because the
results will be unpredictable. The C compiler does not consider it as an
error, though some compilers may give a warning.

Fig. 5.11 The dangling pointer

Therefore, following steps must be followed if pointers are being used:
Steps:
1. Allocate the pointer.
2. Allocate the variable or entity to which the pointer is to be pointed.
3. Point the pointer to the entity.

Let us assume that it is desired that the pointer ptr should point to a
variable val of type int. The correct code in that case would be as
given below:
int*ptr;
intval;
ptr=&val;
*ptr=50;
The result of the above program segment is shown in Figure 5.12.

Fig. 5.12 The correct assignment

5.3 POINTERS AND ARRAYS

Pointers and arrays are very closely related to each other. In C, the
name of an array is a pointer that contains the base address of the array.
In fact, it is a pointer to the first element of the array. Consider the array
declaration given below:
intlist[]={20,30,35,36,39};
This array will be stored in the contiguous locations of the main memory
with starting address equal to 1001(assumed value), as shown in Figure
5.13. Since the array called list is of integer type, each element of this
array occupies two bytes. The name of the array contains the starting
address of the array, i.e., 1001.

Fig. 5.13 The contiguous storage allocation with starting address = 1001
Consider the following program segment:

printf(\nAddressofzerothelementofarraylist=
%d,list);
printf(\nValueofzerothelementofarraylist=
%d,*list);
Once the above program segment is executed, the output would be as
shown below:
Addressofzerothelementofarraylist=1001
Valueofzerothelementofarraylist=20
We could have achieved the same output by the following program
segment also:
printf(\nAddressofzerothelementofarraylist=
%d,&list[0]);
printf(\nValueofzerothelementofarraylist=
%d,list[]);
Both the above given approaches are equivalent because of the following
equivalence relations:
list?&list[0]bothdenoteaddressofzeroth
elementofthearraylist
*list?list[0]bothdenotevalueofzerothelement
ofthearraylist
The left side approach is known as pointer method and right side as
array indexing method. Let us now write a program that prints out a list
by array indexing method.
/*Arrayindexingmethod*/
#include<stdio.h>
main()
{
staticintlist[]={20,30,35,36,39};
inti;
printf(\nThelistis);
for(i=0;i<5;i++)
printf(\n%d%delement,list[i],i);
}
The output of this program would be:

This list is...

20

0element

30

1element

35

2element

36

3element

39

4element

The above program can also be written by pointer method as shown

below:
/*Pointermethodofprocessinganarray*/
#include<stdio.h>
main()
{
staticintlist[]={20,30,35,36,39};
inti;
printf(\nThelistis);
for(i=0;i<5;i++)
printf(\n%d%delement,*(list+i),i);
}
It may be noted in the above program that we have used the term
* (list+i) to print an ith element of the array. Let us analyse this
term. Since list designates the address of zeroth element of the array, we
can access its value through value at address operator, i.e., *. The
following terms are equivalent:
>*list?*(list+0)?list[0]
>*(list+1)?list[1]andsoon
Thus, we can refer to ith element of array list by either of the following
ways:
*(list+i)or*(i+list)orlist[i]
So far, we have used the name of an array to get its base address and
manipulate it. However, there is a small difference between an ordinary
pointer and the name of an array. The difference is that the array name is
a pointer constant and its contents cannot be changed. Thus, the

following operations on list are illegal because they try to change the
contents of a list, a pointer constant.

List=NULL;

/*Notallowed*/

List=&Val;

/*Notallowed*/

list++

/*Notallowed*/

list

/*Notallowed*/

Consider the following program:

#include<stdio.h>
main()
{
chartext[6]=Helpme;
char*p=Helpme;
printf(\n%c,text[3]);
printf(\n%c,p[3]);
}
The output of above program would be:
p
p
From the above program, it looks as if the following declarations are
equivalent as text[3] and p[3]behave in identical manner.
chartext[6];
char*p;
It may be noted that the above declarations are not at all equivalent
though the behaviour may be same. In fact, the array declaration
text [6] asks from the complier for six locations of char type whereas
the pointer declaration char *p asks for a pointer that can point to any
variable of following type:
1. char a character
2. string a string of a characters
3. Null nowhere

Consider the following declarations:

intlist={20,30,35,36,39};
int*p;/*pointervariablep*/
p=list;/*assignthestartingaddressofarray
listtopointerp*/
Since p is a pointer variable and has been assigned the starting address
of array list, the following operations become perfectly valid on this
pointer:
p++,petc.
Let us now write a third version of the program that prints out the array.
We will use pointer variable p in this program.
/*Pointervariablemethodofprocessinganarray*/
#include<stdio.h>
main()
{
staticintlist[]={20,30,35,36,39};
int*p;
inti=0;
p=list;/*Assignthestartingaddressofthelist
*/
printf(\nThelistis);
while(i<5)
{
printf(\n%d%delement,*p,i);
i++;
p++;/*incrementpointer*/
}
}
The output of the above program would be as shown below:

20

0element

30

1element

35

2element

36

3element

39

4element

From our discussion on arrays, we know that the strings are stored and
manipulated as array of characters with last character being a null
character (i.e., \0).
For example, the string ENGINEERING can be stored in an array (say
text) as shown in Figure 5.14.

Fig. 5.14 The array called text

We can declare this string as a normal array by the following array
declaration:
chartext[11];
Consider the following declaration:
char*p;
p=text;
In the above set of statements, we have declared a pointer p that can
point to a character or string of characters. The next statement assigns
the starting address of character string text to the variable
pointer p (see Figure 5.15).

Fig. 5.15 Pointer p points to the array called text

Since text is a pointer constant, its contents cannot be changed. On the
contrary, p is a pointer variable and can be manipulated like any other
pointer as shown in the program given below:
/*Thisprogramillustratestheusageofpointertoa
string*/
#include<stdio.h>
main()
{
chartext[]=ENGINEERING;/*Thestring*/

char*p;/*Thepointer*/
p=text;/*Assignthestartingaddressofstring
top*/
printf(\nThestring..);/*Printthestring*/
while(*p!=\0)
{printf(%c,*p);
p++;
}
}
Example 2: What would be the output of the following code?
#include<stdio.h>
{
char*ptr;
ptr=Nice;
printf(\n%c,*&ptr[1]);
}
Solution: The output would be: i.
Example 3: What would be the output of the following code?

#include<stdio.h>
main()
{
intlist[5]={2,5,6,0,9};
3[list]=4[list]+6;
printf(%d,*(list+3));
}
Solution: The output would be 15.
5.4 ARRAY OF POINTERS

Like any other array, we can also have an array of pointers. Each element
of such an array can point to a data item such as variable, array etc. For
example, consider the declaration given below:
float*x[20];

This declaration means that x is an array of 20 elements and each

element is a pointer to a variable of type float. Let us now construct an
array of pointers to strings.
char*item[]{Chair,
Table,
Stool,
Desk,
};
The above declaration gives an array of pointers called item. Each
element of item points to a string as shown in Figure 5.16.

Fig. 5.16 Array of pointers to strings

The advantage of having an array of pointer to strings is that
manipulation of strings becomes convenient. For example, the string
containing table can be copied into a pointer ptr without
usingstrcpy() function by the following set of statements:
char*ptr;//declareapointertoastring
ptr=item[1];//assigntheappropriatepointerto
ptr
It may be noted that once the above set of statements is executed, both
the pointers item[1] and ptrwill point to the same string i.e., table as
shown in Figure 5.17.

Fig. 5.17 Effect of statement ptr = item[1]

However, the changes made by the pointer ptr to the string table will
definitely disturb the contents of the string pointed by the pointer item
[1] and vice versa. Therefore, utmost care should be taken by the
programmer while manipulating pointers pointing to the same memory
locations.
5.5 POINTERS AND STRUCTURES

We can have a pointer to a structure, i.e., we can also point a pointer to a

structure. A pointer to a structure variable can be declared by prefixing
the variable name by a *. Consider the following declaration:
structxyz{/*structuredeclaration*/
inta;
floaty;
};
structxyzabc;/*declareavariableoftypexyz
*/
structxyz*ptr;/*declareapointerptrtoa
variableoftypexyz*/
Once the above declarations are made, we can assign the address of the
structure variable abc to the structure pointer ptr by the following
statement:
ptr=&abc;

Since ptr is a pointer to the structure variable abc, the members of the
structure can also be accessed through a special operator called arrow
operator, i.e., (minus sign followed by greater than sign). For
example, the members a and y of the structure variable abc pointed
by ptr can be assigned values 30 and 50.9, respectively by the following
statement:
ptr>a=30;
ptr>y=50.9;
The arrowoperator> is also known as a pointertomember
operator.
Example 4: Write a program that defines a structure called item for
the record structure given below. It reads the data of the record through
dot operator and prints the record by arrowoperator. Use suitable
data types for the members of the structure called item.
Solution: We will use a pointer ptr to point to the structure called
item. The required program is given below:
/*Thisprogramdemonstratestheusageofanarrow
operator*/
#include<stdio.h>
main()
{
structitem{
charcode[5];
intQty;
floatcost;
};
structitemitem_rec;/*Defineavariableofstruct
type*/
structitem*ptr;/*Defineapointeroftypestruct
*/
/*Readdatathroughdotoperator*/
printf(\nEnterthedataforanitem);
printf(\nCode:);scanf(%s,&item_rec.code);
printf(\nQty:);scanf(%d,&item_rec.Qty);
printf(\nCost:);scanf(%f,&item_rec.cost);
/*Assigntheaddressofitem_rec*/

ptr=&item_rec;
/*Printdatathrougharrowoperator*/
printf(\nThedatafortheitem);
printf(\nCode:%s,ptr>code);
printf(\nQty:%d,ptr>Qty);
printf(\nCost:%5.2f,ptr>cost);
}
From the above program, we can see that the members of a static
structure can be accessed by both dot and arrow operators. However, dot
operator is used for simple variable whereas the arrow operator is used
for pointer variables.
Example 5: What is the output of the following program?
#include<stdio.h>
main()
{
structpoint
{
intx,y;
}polygon[]={{1,2},{1,4},{2,4},{2,2}};
structpoint*ptr;
ptr=polygon;
ptr++;
ptr>x++;
printf(\n%d,ptr>x);
}
Solution: From the above program, it can be observed that ptr is a
pointer to an array of structures called polygon. The statement ptr+
+ moves the pointer from zeroth location (i.e., {1, 2}) to first location
(i.e., {1, 4}). Now, the statement x++ has incremented the field x of the
structure by one (i.e., 1 has incremented to 2).
The output would be 2.
5.6 DYNAMIC ALLOCATION

We know that in a program when a variable x is declared, a storage

location is ade available to the program as shown in Figure 5.18.
This memory location remains available, even if it is not needed, to the
program as long as the program is being executed. Therefore, the

variable x is known as a static variable. Dynamic variables, on the

other hand, can be allocated from or returned to the system according to
the needs of a running program. However, a dynamic variable does not
have a name and can be referenced only through a pointer.

Fig. 5.18 Memory allocation to variable x

Consider the declaration given below:
int*p;
The above declaration gives us a dangling pointer p that points to
nowhere (see Figure 5.19). Now, we can connect this dangling pointer
either to a static variable or a dynamic variable. In our previous
discussion on pointers, we have always used pointers with static
variables. Let us now see how the pointer can be connected to a dynamic
variable.

Fig. 5.19 The dangling pointer p

In C, dynamic memory can be allocated by a function called malloc().
The format of this function is given below:
(<type>*)(malloc(<size>))
wher
e

<type:
>

is the type of variable, which is required at run

time.

malloc
:

is a reserved word indicating that memory

allocation is to be done.

<size>
:

is the size of memory to be allocated. In fact, the

size of any variable can be obtained
by sizeof() function.

Consider the following declaration:

int*p;intSz;
Sz=sizeof(int);

P=(int*)malloc(Sz);
The above set of statements is executed in the following manner;
1. The pointer p is created. It is dangling, i.e., it is not connected anywhere
(see Figure 5.20 (a))
2. A variable called Sz is declared to store the size of dynamic variable.
3. The size of dynamic variable (int in this case) is stored in the variable
called Sz.
4. A dynamic variable of type integer is taken through malloc() function and
the dangling pointerp is connected to the variable (see Figure 5.20 (b)).

It may be noted from Figure 5.20 (b) that a dynamic variable is

anonymous in the sense that it has no name and, therefore, it can be
referenced only through the pointer p. For example, a value (say 50) can
be assigned to the dynamic variable by the following statement:

Fig. 5.20 The effect of malloc() function

*p=50;
The above statement means that it is desired to store a value 50 in a
memory location pointed to by the pointer p. The effect of this
statement is shown in Figure 5.21.
Fig. 5.21 The effect of assignment of *p = 50
It may be observed that p is simply a pointer or an address and *p gives
the access to the memory location or the variable where it is pointing to.
It is very important to learn the difference between the pointer and its
associated dynamic variable. Let us consider the following program
segment:
int*Aptr,*Bptr;/*declarepointers*/
Aptr=(int*)malloc(sizeofint));/*allocate
dynamicmemory*/
Bptr=(int*)malloc(sizeof(int));
*Aptr=15;

*Bptr=70;
The outcome of the above program segment would be as shown in Figure
5.22.

Fig. 5.22 Value assignment through pointers

Let us now see the effect of the following assignment statement on the
memory allocated to the pointers AptrandBptr.
Aptr=Bptr;
The above assignment makes Aptr point to the same location or variable
which is already being pointed by Bptr as shown in Figure 5.23. Thus,
both Aptr and Bptr are now pointing to the same memory location.
A close observation would reveal that the dynamic variable containing
value 15 is now lost in the sense that it is neither available to Aptr any
more and nor it is available to the system. The reason is that the system
gave it to the pointer Aptr and now Aptr is pointing to some other
location. This dynamic variable, being anonymous, (i.e., without name)
is no more accessible. Even if we want to reverse the process, we cannot
simply do it. This phenomenon of loosing dynamic variables is known
as memory bleeding. Thus, utmost care should be taken while
manipulating the dynamic variables. It is better to return a dynamic
variable whenever it is not required in the program.

Fig. 5.23 Both pointers pointing to the same location

A dynamic variable can be returned to the system by function
called free(). The general form of usage of this operator is given
below:
free(p_var);
wher

free:

is a reserved word

e
p_var
:

is the pointer to the dynamic variable to be returned

to the system.

Let us consider pointers shown in Figure 5.23. If it is desired to return

the dynamic variable pointed by pointer Bptr, we can do so by the
following statement:
free(Bptr);
Once the above statement is executed, the dynamic variable is returned
back to the system and we are left with two dangling
pointers Bptr and Aptr as shown in Figure 5.24. The reason for this is
obvious because both were pointing to the same location. However, the
dynamic variable containing value 15 is anonymous and not available
anymore and, hence, a total waste of memory.

Fig. 5.24 The two dangling pointers and a lost dynamic memory
variable
Example 6: Write a program that dynamically allocates an integer. It
initializes the integer with a value, increments it, and print the
incremented value.

Solution: The required program is given below:

#include<stdio.h>
#include<alloc.h>
main()
{
int*p;/*Apointertoanint*/
intSz;
/*Computethesizeoftheint*/
Sz=sizeof(int);
/*Allocateadynamicintpointedbyp*/

p=(int*)malloc(Sz);
printf(\nEnteravalue:);
scanf(%d,p);
*p=*p+1;
/*Printtheincrementedvalue*/
printf(\nTheValue=%d,*p);
free(p);
}
Example 7: Write a program that dynamically allocates a structure
whose structure diagram is given below. It reads the various members of
the structure and prints them.
Solution: We will use the following steps to write the required program:
1.
2.
3.
4.
5.
6.
7.

Declare a structure called student.

Declare a pointer ptr to the structure student.
Compute the size of the structure.
Ask for dynamic memory of type student pointed by the pointer ptr.
Read the data of the various elements using arrow operator.
Print the data.
Return the dynamic memory pointed by ptr.

The required program is given below:

#include<stdio.h>
#include<alloc.h>
main()
{
structstudent{/*Declarethestructure*/
charname[15];
intage;
introll;
};
structstudent*ptr;
intSz;
/*Computethesizeofstruct*/
Sz=sizeof(structstudent);
/*Askfordynamicmemoryoftypestudent*/
ptr=(structstudent*)malloc(Sz);
printf(\nEnterthedataofthestudent);

printf(\nName:);/*fflush(stdin);*/gets(ptr>
name);
printf(\nAge:);scanf(%d,&ptr>age);
printf(\nRoll:);scanf(%d,&ptr>roll);
printf(\nThedatais...);
printf(\nName:);puts(ptr>name);
printf(\nAge:%d,ptr>age);
printf(\nRoll:%d,ptr>roll);
free(ptr);
}
Note: An array can also be dynamically allocated as demonstrated in the
following example.
Example 8: Write a program that dynamically allocates an array of
integers. A list of integers is read from the keyboard and stored in the
array. The program determines the smallest in the list and prints its
location in the list.
Solution: The solution to this problem is trivial and the required
program is given below:

/*Thisprogramillustratestheusageofdynamically
allocatedarray*/
#include<stdio.h>
main()
{
inti,min,pos;
int*list;
intSz,N;
printf(\nEnterthesizeofthelist:);
scanf(%d,&N);
Sz=sizeof(int)*N;/*Computethesizeofthe
list*/
/*AllocatedynamicarraysizeN*/
list=(int*)malloc(Sz);
printf(\nEnterthelist);
/*Readthelist*/
for(i=0;i<N;i++)
{

printf(\nEnterNumber:);
scanf(%d,(list+i));
}
pos=0;/*Assumethatthezerothelementismin*/
min=*(list+0);
/*Findtheminimum*/
for(i=1;i<N;i++)
{
if(min>*(list+i))
{min=*(list+i);
pos=i;
}
}/*Printtheminimumanditslocation*/
printf(\nTheminimumis=%datposition=%d,
min,pos);
free(list);
}
From the above example, it can be observed that the size of a
dynamically allocated array can be specified even at the run time and the
required amount of memory is allocated. This is in sharp contrast to
static arrays for whom the size have to be declared at the compile time.
5.6.1 Self Referential Structures
When a member of a structure is declared as a pointer to the structure
itself then the structure is called a self referential structure. Consider the
following declaration:
structchain{
intval;
structchain*p;
};
The structure chain consists of two members: val and p. The member
val is a variable of type intwhereas the member p is a pointer to a
structure of type chain. Thus, the structure chain has a member that can
point to a structure of type chain or may be itself. This type of self
referencing structure can be viewed as shown in Figure 5.25.

Fig. 5.25 Self referential structure chain

Since pointer p can point to a structure variable of type chain, we can

connect two such structure variables A and B to obtain a linked structure
as shown in Figure 5.26.

Fig. 5.26 Linked structure

The linked structure given in Figure 5.26 can be obtained by the
following steps:
1. Declare structure chain.
2. Declare variable A and B of type chain.
3. Assign the address of structure B to member p of structure A.

These above steps have been coded in the program segment given below:
structchain{/*Declarestructurechain*/
intval;
chain*p;
};
structchainA,B;/*DeclarestructurevariablesA
andB*/
A.p=&B;/*ConnectAtoB*/
From Figure 5.26 and the above program segment, we observe that the
pointer p of structure variable Bis dangling, i.e., it is pointing to
nowhere. Such pointer can be assigned to NULL, a constant indicating
that there is no valid address in this pointer. The following statement
will do the desired operation:
B.p=NULL;
The data elements in this linked structure can be assigned by the
following statements:
A.val=50;
B.val=60;
The linked structure now looks like as shown in Figure 5.27.

Fig. 5.27 Value assignment to data elements

We can see that the members of structure B can be reached by two

methods:
1. From its variable name B through dot operator.
2. From the pointer p of variable A because it is also pointing to the
structure B. However, in this case the arrow operator is needed.

Consider the statements given below:

printf(\nthecontentsofmembervalofB=%d,
B.val);
printf(\nthecontentsofmembervalofB=%d,
A.p>val);
Once the above statements are executed, the output would be:
ThecontentsofmembervalofB=60
ThecontentsofmembervalofB=60
The linked structures have great potential and can be used in numerous
programming situations such as lists, trees, etc.
Example 9: Write a program that uses self referential structures to
create a linked list of nodes of following structure. While creating the
linked list, the student data is read. The list is also travelled to print the
data.

Solution: We will use the following self referential structure for the
purpose of creating a node of the linked list.
structstud{
charname[15];
introll;
structstudnext;
};
The required linked list would be created by using three
pointers: first,farandback. The following algorithm would be
employed to create the list:
Step
1. Take a new node in pointer called first.

2. Read first name and first roll.

3. Point back pointer to the same node being pointed by first, i.e., back 5 first.
4. Bring a new node in the pointer called far.
5. Read far name and far roll.
6. Connect next of back to for, i.e., back next 5 far.
7. Take back to far, i.e., back 5 far.
8. Repeat steps 4 to 7 till whole of the list is constructed.
9. Point next of far to NULL, i.e., far next = NULL.
10. Stop.

Fig. 5.28 Creation of a linked list

The required program is given below:
#include<stdio.h>
#include<alloc.h>
main()
{
structstudent{
charname[15];
introll;
structstudent*next;
};
structstudent*First,*Far,*back;
intN,i;
intSz;

Sz=sizeof(structstudent);
printf(\nEnterthenumberofstudentsinthe
class);
scanf(%d,&N);
/*Takefirstnode*/
First=(structstudent*)malloc(Sz);
/*Readthedataofthefirststudent*/
printf(\nEnterthedata);
printf(\nName:);fflush(stdin);gets(First>
name);
printf(\nRoll:);scanf(%d,&First>roll);
/*pointbackwhereFirstpoints*/
back=First;
for(i=2;i<=N;i++)
{/*BringanewnodeinFar*/
Far=(structstudent*)malloc(Sz);
/*ReadDataofnextstudent*/
printf(\nName:);fflush(stdin);gets(Far>
name);
printf(\nRoll:);scanf(%d,&Far>roll);
/*ConnectBacktoFar*/
back>next=Far;
/*pointbackwhereFarpoints*/
back=Far;
}/*Repeattheprocess*/
Far>next=NULL;
/*Printthecreatedlinkedlist*/
Far=First;/*Pointtothefirstnode*/
printf(\nTheDatais);
while(Far!=NULL)
{
printf(\nName:%s,Roll:%d,Far>name,Far
>roll);
/*Pointtothenextnode*/
Far=Far>next;
}
}
Example 10: Modify the program developed in Example 1 such that
1. assume that the number of students in the list is not known.

2. the list of students is split into the two sublists pointed by two
pointers: front_half andback_half. The front_half points to the
front half of the list and the back_half to the back half. If the number of
students is odd, the extra student should go in the front list.

Solution: The list of students would be created in the same fashion as

done in Example1. The following steps would be followed to split the
list in the desired manner:
1. Travel the list to count the number of students.
2. Compute the middle of the list.
3. Point the pointer called front_half to first and travel the list again
starting from the first and reach to the middle.
4. Point a pointer called backhalf to the node that succeeds the middle of
the list.
5. Attach NULL to the next pointer of the middle node.
6. Print the two lists, i.e., pointed by fronthalf and backhalf.

The required program is given below:

#include<stdio.h>
#include<alloc.h>
main()
{
structstudent{
charname[15];
introll;
structstudent*next;
};
structstudent*First,*Far,*back;
structstudent*front_half,*back_half;
intN,i,count,middle;
intSz;
Sz=sizeof(structstudent);
printf(\nEnterthenumberofstudentsinthe
class);
scanf(%d,&N);
/*Takefirstnode*/
First=(structstudent*)malloc(Sz);
/*Readthedataofthefirststudent*/
printf(\nEnterthedata);
printf(\nName:);fflush(stdin);gets(First>
name);

printf(\nRoll:);scanf(%d,&First>roll);
First>next=NULL;
/*pointbackwhereFirstpoints*/
back=First;
for(i=2;i<=N;i++)
{/*BringanewnodeinFar*/
Far=(structstudent*)malloc(Sz);
/*Readdataofnextstudent*/
printf(\nName:);fflush(stdin);gets(Far>
name);
printf(\nRoll:);scanf(%d,&Far>roll);
/*ConnectBacktoFar*/
back>next=Far;
/*PointbackwhereFarpoints*/
back=Far;
}/*Repeattheprocess*/
Far>next=NULL;
/*Countthenumberofnodes*/
Far=First;/*Pointtothefirstnode*/
count=0;
while(Far!=NULL)
{
count++;
Far=Far>next;
}/*Splitthelist*/
if(count==1)
{
printf(\nThelistcannotbesplit);
front_half=First;
}
else
{
/*Computethemiddle*/
if((count%2)==0)middle=count/2;
elsemiddle=count/2+1;
/*Travelthelistandsplitit*/
front_half=First;
Far=First;
for(i=1;i<=middle;i++)
{
back=Far;

Far=Far>next;
}
back_half=Far;
back>next=NULL;
}
/*Printthetwolists*/
Far=front_half;
printf(\nTheFrontHalf:);
for(i=1;i<=2;i++)
{
while(Far!=NULL)
{
printf(Name:%s,Roll:%d,Far>name,Far>
roll);
Far=Far>next;
}
if(count==1||i==2)break;
printf(\nTheBackHalf:);
Far=back_half;
}
}
FREQUENTLY ASKED QUESTIONS
1. What is meant by a pointer?
Ans. It is a variable which can only store the address of another variable.
2. Are the expressions (*p)++and++*p equivalent?
Ans. Yes
3. What happens if a programmer assigns a new address to a dynamic variable
before he either assigns its address to another pointer or deletes the
variable?
Ans. The dynamic variable will be lost, i.e., it becomes inaccessible to
program as well as the operating system.
4. What is meant by memory bleeding?
Ans. The lost dynamic variables lead to memory bleeding.
5. Pointers always contain integers. Comment.
Ans. We know that a pointer contains an address of other variables. Since
an address can only be a whole number or integer, the pointers contain
integers only.
6. What does the term **j mean?
Ans. The term **j means that j is a pointer to a pointer.
7. Are the expressions *p++and++*p equivalent?

Ans. No. The expression *p++ returns the contents of the variable pointed
by p and then increments the pointer. On the other hand, the expression +
+*p returns the incremented contents of the variable pointed by p.
8. Differentiate between an uninitialized pointer and a NULL pointer.
Ans. An un initialized pointer is a dangling pointer which may contain any
erroneous address. ANULL pointer does not point to any address location.

TEST YOUR SKILLS

1. Give the output of the following program:
#include<stdio.h>
main()
{
inti,j;
char*str=CALIFORNIA;
for(i=0;str[i];i++)
{
for(j=0;j<=i;j++)
printf(%c,str[j]);
printf(\n);
}
}
Ans. The output would be
C
CA
CAL
CALI
CALIF
CALIFO
CALIFOR
CALIFORN
CALIFORNI
CALIFORNIA
2. What would be the output of the following?
#include<stdio.h>
main()
{
inta;
*&a=50;
printf(%d,a);
}

Ans. The output would be 50.

3. What would be the output of the following code?
#include<stdio.h>
main()
{inti=50;
int*j=&i;
printf(\n%d,++*(j));
}
Ans. The output would be 51.
4. What would be the output of the following?
#include<stdio.h>
main()
{
char*ptr=abcd;
charch=++*ptr++;
printf(\n%c,ch);
}
Ans. The output would be b.
5. What would be the output of the following code? Assume that the array
starts at location 5714 in the memory.
#include<stdio.h>
main()
{
inttab[3][4]={5,6,7,8,
1,2,3,4,
9,10,0,11};
printf(\n%d%d,*tab[0]+1,*(tab[0]+1));
printf(\n%d,*(*(tab+0)+1));
}
Ans. The output would be
66
6
6. What would be the output of the following code?

#include<stdio.h>
main()
{
structval{
intNet;
};
structvalx;
structval*p;
p=&x;
p>Net=50;
printf(\n%d,++(x.Net));
}
Ans. The output would be 50.
7. What would be the output of the following code?

#include<stdio.h>
main()
{
structval{
intNet;
};
structvalx;
structval*p;
p=&x;
p>Net=50;
printf(\n%d,(x.Net)++);
}
Ans. The output would be 21.

EXERCISES
1. Explain the & and * operators in detail with suitable examples.
2. Find the errors in the following program segments:

a.
intval=10
int*p;
p=val;
b.
charlist[10];
charp;
list=p;
2. (i) What will be the output of the following program:

#include<stdio.h>
main()
{
intval=10;
int*p,**k;
p=&val;
k=&p;
printf(\n%d%d%d%d,p,*p,*k,**k);
}
(ii)
#include<stdio.h>
main()
{
charch;
char*p;
ch=A;
p=&ch;
printf(\n%c%c,ch,(*p)++);
}
3. What will be the output of the following:
#include<stdio.h>
main()
{
intlist[5],i;
*list=5;
for(i=1;i<5;i++)
*(list+i)=*(list+i1)*i;
printf(\n);
for(i=0;i<5;i++)
printf(%d,*(list+i));
}
4. Find the errors in the following program segment:
structxyz{
char*A;
char*B;
intval;
};
xyzs;
A=Firsttext;

B=Secondtext;
5. What will be the output of the following program segment?
#include<stdio.h>
main()
{
structMyrec{
charch;
structMyrec*link;
};
structMyrecx,y,z,*p;
x.ch=x;
y.ch=y;
z.ch=z;
x.link=&z;
y.link=&x;
z.link=NULL;
p=&y;
while(p!=NULL)
{
printf(%c,p>ch);
p=p>link;
}
}
6. What will be the output of the following program?
#include<stdio.h>
main()
{
floatcost[]={35.2,37.2,35.42,36.3};
float*ptr[4];
inti;
for(i=0;i<4;i++)
*(ptr+i)=cost+i;
for(i=0;i<4;i++)
printf(%f,*(*(ptr+i)));
}
Assume the base address of the array cost to be 2025.
7. What will be the output of the following program:
#include<stdio.h>

main()
{
charitem[]=COMPUTER;
inti;
for(i=7;i>=0;i)
printf(%c,*(item+i));
}

6
Linked Lists
CHAPTER OUTLINE

6.1 Introduction
6.2 Linked Lists
6.3 Operations on Linked Lists
6.4 Variations of Linked Lists
6.5 The Concept of Dummy Nodes
6.6 Linked Stacks
6.7 Linked Queues
6.8 Comparison of Sequential and Linked Storage
6.9 Solved Problems
6.1 INTRODUCTION

An array is a very simple and extremely useful data structure. A large

number of applications based on stacks and queues can be easily
implemented with the help of arrays. Lists and tables have natural
similarity with one and two-dimensional arrays. Though the allocation of
space is sequential in an array, the access to its elements is random. For
a given index the element can be accessed in a time, independent of its
location in the array.
However, there are many problems associated with this data structure.
Some of the important problems are as follows:
1. An array is a static data structure and, therefore, its size should be known
in advance before its usage.
2. If a list of data items is to be stored, then an array of approximate size is
used, i.e., the array has to be large enough to hold the maximum amount
of data one could logically expect. This is totally a guess work and can
result in overflow or wastage of main memory.
3. What if the number of elements in the list is not known in advance?
4. Insertion or deletion operations in an array, except from the end, are a
costly exercise. These operations require large number of shifting of
elements in one direction or the other. This a very time consuming exercise.

The above problems can be dealt with the help of linked structures,
discussed in the subsequent sections.
6.2 LINKED LISTS

Linked structures are a completely different way to represent a list of

items. Each element of the list is made to point to the next element in
the list as shown in Figure 6.1(a).
An element in the list is called a node. The node is a selfreferential structure, having two parts: Dataand Next as shown in Figure
6.1(b). The Data part contains the information about the element and
the Next part contains a pointer to the next element or node in the list.
For example, in Figure 6.1(a), a list consisting four names (Preksha,
Sagun, Bhawna, and Ridhi) is represented using linked structures.
Such a list is called linked list or a linear linked list or a singly linked list.
Linked list is a series of nodes. Each node has two parts: dataand
next. The data part contains the information about the node and
the next part is a pointer that points to nextnode. The next of
last node points to NULL.

Fig. 6.1 Linked representation of a list

It may be noted that the list is pointed by a pointer called List.
Currently List is pointing to a nodecontaining the name Preksha, the
next part of which is pointing to Sagun, the next of which is pointing to
Bhawna, and so on.

Fig. 6.2 A node called X

By convention, a nodeX means that the node is being pointed by a
pointer X as shown in Figure 6.2. In an algorithm, the data part of X is
written as DATA(X) and the next part is written as NEXT(X).

Let us now insert a new node containing data called Samridhi between
Preksha and Sagun. This insertion operation can be carried out by
the following steps:
Step
1. Getanodecalledptr.
2. DATA(ptr)=Samridhi
3. NEXT(ptr)=NEXT(List)
4. NEXT(List)=ptr

The third step means that the Next pointer of ptr should point to the
same location that is currently being pointed by Next of List.
The effect of above steps (14) is shown in Figure 6.3.

Fig. 6.3 Insertion of a node in the linked list

It may be noted that the insertion operation is a very simple and easy
operation in a linked list because, unlike arrays, no copying or shifting
operations are required.
Similarly, it is easy to delete an element from a linked list. To delete
Bhawna from the list given inFigure 6.3, we need only to redirect the

next pointer of Sagun to point to Ridhi as shown in Figure 6.4.This can

be done by the following the following steps:
Step
1. Pointapointercalledptr1tothenode
containingthenameBhawna.
2. Pointapointercalledptr2tothenodecontainingthe
nameSagun.
3. Next(ptr2)=Next(ptr1)
4. Returnptr1

The effect of the above code segment is shown in Figure 6.4.

Fig. 6.4 Deletion of a node

From the above discussion, it is clear that unlike arrays, there is no
longer a direct correspondence between the logical order of elements
present in a list and the way the nodes are arranged physically in
the linkedlist (see Figure 6.5).

Fig. 6.5 Logical versus physical order in a linked list

The physical order of the nodes shown in Figure 6.5 is I My Love India
whereas the logical order is I Love My India. It is because of this
feature that a linked list offers a very easy way of insertions and deletions
in a list without the physical shifting of elements in the list. While

manipulating a linkedlist, the programmer keeps in mind only the

logical order of the elements in the list. Physically, how the nodes are
stored in the memory is neither known to nor the concern of the
programmer.
6.3 OPERATIONS ON LINKED LISTS

The linkedlist is a dynamic datastructure. The following

operations are associated withlinkedlists:
1.
2.
3.
4.
5.

Creation of a linkedlists
Traversing a linkedlists
Searching in a linkedlists
Inserting a node in a linkedlists
Deleting a node from a linkedlists

6.3.1 Creation of a Linked List

From the above discussion, we know that a node of a linkedlist is
essentially a structure because it contains data of different types. In
addition to the information part, it contains a pointer that can point to a
node, i.e., to itself or to some other node. Such structures are called selfreferential structures.
6.3.1.1 Self-referential Structures
When a member of a structure is declared as a pointer to the structure
itself, then the structure is called self-referential structure. Consider the
following declaration:
structchain{intval;
structchain*p;
};
The structure called chain consists of two members: val and p. The
member val is a variable of type int whereas the member p is a pointer
to a structure of type chain. Thus, the structure chain has a member that
can point to a structure of type chain or may be itself. This type of selfreferencing structure can be viewed as shown in Figure 6.6.
Fig. 6.6 Self-referential structure chain

Fig. 6.7 Linked structure

Since pointer p can point to a structure variable of type chain, we can
connect two such structure variables, A and B, to obtain a linked
structure as shown in Figure 6.7.
The linked structure given in Figure 6.7 can be obtained by the following
steps:
1. Declare structurechain.
2. Declare variables A and B of type chain.
3. p(A)=B

These steps have been coded in the program segment given below:
structchain{/*declarestructurechain*/
intval;
chain*p;
};
structchainA,B;/*declarestructurevariablesA
andB*/
A.p=&B;/*connectAtoB*/
From Figure 6.7 and the above program segment, we observe that the
pointer p of structure variable Bis dangling, i.e., it is pointing to
nowhere. Such pointer can be assigned to NULL, a constant indicating
that there is no valid address in this pointer. The following statement
will do the desired operation:
B.p=NULL;
The data elements in this linked structure can be assigned by the
following statements:
Fig. 6.8 Value assignment to data elements
A.val=50;
B.val=60;
The linked structure now looks like as shown in Figure 6.8.
We can see that the members of structure B can be reached by the
following two methods:

1. From its variable name B through dot operator.

2. From the pointer pofvariableA because it is also pointing to the
structure B. However, in this case the arrow operator (->) is needed to
access the field called val as shown below.

Consider the statements given below:

printf(\nthecontentsofmembervalofB=%d,
B.val);
printf(\nthecontentsofmembervalofB=%d,A
>p.val);
Once the above statements are executed, the output would be:
ThecontentsofmembervalofB=60
ThecontentsofmembervalofB=60
The linked structures have great potential and can be used in numerous
programming situations such as lists, trees, etc. The major advantage of
linked structures is that the pointers can be used to allocate dynamic
memory. Consider the following declaration:
structchain*ptr;
The above statement declares a pointer called ptr, which can point to a
structure of type chain [seeFigure 6.9 (a)]. Lets us use ptr to take one
such structure through dynamic allocation with the help
ofmalloc() function of C. From Chapter 5, we know that
the malloc() function needs the size of the memory that is desired by
the programmer. Therefore, in the first step, given below, the size of the
structure has been obtained and in the second step the memory of that
size has been asked throughmalloc() function.

Fig. 6.9 Dynamic allocation of a structure

Size=sizeof(structchain);
ptr=(structstudent*)malloc(Size);

The effect of the above statements is shown in Figure 6.9 (b). It may be
noted that the node pointed byptr has no name. In fact all dynamic
memory allocations are anonymous.
Let us now use self-referential structures to create a linked list of nodes
of structure type given inFigure 6.10. While creating the
linked list, the student data is read. The list is also travelled to print
the data.
Fig. 6.10 The self-referential structure student
We will use the following self-referential structure for the purpose of
creating a node of the linked list:
structstud{
charname[15];
introll;
structstud*next;
};
The required linked list would be created by using three pointers
first,far, and back. The following algorithm would be employed to
create the list:
Algorithm createLinkList()
{
Step
1.Takeanewnodeinpointercalledfirst.
2.Readfirst>nameandfirst>roll.
3.Pointbackpointertothesamenodebeingpointed
byfirst,i.e.,back=first.
4.Bringanewnodeinthepointercalledfar.
5.Readfar>nameandfar>roll.
6.Connectnextofbacktofar,i.e.,back>next=
far.
7.Takebacktofar,i.e.,back=far.
8.Repeatsteps4to7tillwholeofthelistis
constructed.
9.PointnextoffartoNULL,i.e.,far>next=
NULL.
10.Stop.
}

The effect of the above steps is shown in Figure 6.11.

The required program is given below:
#include<stdio.h>
#include<alloc.h>
structstudent{
charname[15];

Fig. 6.11 Creation of a linked list

introll;
structstudent*next;
};
structstudent*create();/*Functionthat
createsthelinkedlist*/
voiddispList(structstudent*First);/*
Functiontodisplaylinkedlist*/
voidmain()
{
structstudent*First;
First=create();
dispList(First);
}
structstudent*create()
{

structstudent*First,*Far,*back;
intN,i;
intSz;
Sz=sizeof(structstudent);
printf(\nEnterthenumberofstudentsinthe
class);
scanf(%d,&N);
if(N==0)returnNULL;/*Listisempty*/
/*Takefirstnode*/
First=(structstudent*)malloc(Sz);
/*Readthedataofthefirststudent*/
printf(\nEnterthedata);
printf(\nName:);fflush(stdin);gets(First
>name);
printf(\nRoll:);scanf(%d,&First>roll);
First>next=NULL;
/*pointbackwhereFirstpoints*/
back=First;
for(i=2;i<=N;i++)
{/*BringanewnodeinFar*/
Far=(structstudent*)malloc(Sz);
/*ReadDataofnextstudent*/
printf(\nName:);fflush(stdin);gets(Far>name);
printf(\nRoll:);scanf(%d,&Far>roll);
/*ConnectBacktoFar*/
back>next=Far;
/*pointbackwhereFarpoints*/
back=Far;
}/*Repeattheprocess*/
Far>next=NULL;
returnFirst;/*returnthepointertothe
linkedlist*/
}
/*Printthecreatedlinkedlist*/
voiddispList(structstudent*First)
{structstudent*Far;
Far=First;/*Pointtothefirstnode*/

printf(\nTheListis.);
while(Far!=NULL)
{
printf(%s%d,,Far>name,Far>roll);
/*Pointtothenextnode*/
Far=Far>next;
}
}
It may be noted that the next pointer of the last node of a linked list
always points to NULL to indicate the end of the list. The pointer First
points to the beginning of the list. The address of an in between node is
not known and, therefore, whenever it is desired to access a particular
node, the travel to that node has to start from the beginning.
6.3.2 Travelling a Linked List
Many a times, it is required to travel whole or a part of the list. For
example, the operations such as counting of nodes, processing of
information of some nodes, or printing of the list, etc. requires traversal
of the list.
Consider the linked list shown in Figure 6.12. It is pointed by First. Let
us travel this list for the purpose of counting the number of nodes in the
list.

Fig. 6.12 The list pointed by First

Let us use a pointer ptr to travel the list. A variable called count would
be used to count the number of nodes in the list. The travel stops when a
NULL is encountered. The following algorithm
calledtravelList() travels the given list pointed by First.
Algorithm travelList()
{
Step
1.ifFirst==NULLthen{printListEmpty;Stop}
2.ptr=First;/*ptrpointstothenodebeing
pointedbyFirst*/
3.count=0;
4.while(ptr!=Null)
{
4.1count=count+1;

4.2ptr=NEXT(ptr);
}
5.printThenumberofnodes=,count;
6.End
In the above algorithm, step4.2, given below, is worth noting:
ptr=NEXT(ptr)
The above statement says that let ptr point to the same location which is
currently being pointed by NEXT of ptr. The effect of this statement on
the list of Figure 6.12 is shown in Figure 6.13.
It may be noted that the step4.2 takes the ptr to next node. Since
this statement is within the scope of the while loop, ptr travels from the
First to the NULL pointer, i.e., the last node of the list.
Example 1: Write a program that travels a linked list consisting of
nodes of following struct type. While traveling, it counts the number of
nodes in the list. Finally, the count is printed.

structstudent{
charname[15];
introll;.
structstudent*next;
};

Fig. 6.13 Significance of the operation ptr = NEXT (ptr)

Solution: A linked list of nodes of student type would be created by
using a function called create(). The nodes would be added to the list
until the user enters a roll number less than or equal to 0, i.e., (roll <=
0). A function countNode() based on the
algorithm travelList() would be used to travel and count the nodes
in the list.
The required program is given below:

/*Thisprogramtravelsalinkedlistandcountsthe
numberofnodes*/
#include<stdio.h>
#include<alloc.h>
structstudent{
charname[15];
introll;
structstudent*next;
};
structstudent*create();
intcountNode(structstudent*First);
voidmain()
{
intcount;
structstudent*First;
First=create();
count=countNode(First);
printf(\nThenumberofnodes=%d,count);
}
/*Thisfunctioncreatesalinkedlist*/
structstudent*create()
{
structstudent*First,*Far,*back;
inti;
intSz;
Sz=sizeof(structstudent);
/*Takefirstnode*/
First=(structstudent*)malloc(Sz);
/*Readthedataofthefirststudent*/
printf(\nEnterthedataofthefirststudent);
printf(\nName:);fflush(stdin);gets(First
>name);
printf(\nRoll:);scanf(%d,&First>roll);
First>next=NULL;
if(First>roll<=0){
printf(\nNodes=0);

First=NULL;
exit(1);}/*Emptylist*/
/*pointbackwhereFirstpoints*/
back=Far=First;
while(Far>roll>0)
{/*BringanewnodeinFar*/
Far=(structstudent*)malloc(Sz);
/*ReadDataofnextstudent*/
printf(\nName:);fflush(stdin);gets(Far
>name);
printf(\nRoll:);scanf(%d,&Far>roll);
if(Far>roll<=0)break;/*Endofthelist*/
/*ConnectBacktoFar*/
back>next=Far;
/*pointbackwhereFarpoints*/
back=Far;
}/*repeattheprocess*/
back>next=NULL;
returnFirst;
}
/*Travelandcountthenumberof
nodes*/
intcountNode(structstudent*First)
{
structstudent*ptr;
intcount;
count=0;
ptr=First;/*Pointtothefirstnode*/
while(ptr!=NULL)
{count=count+1;
/*Pointtothenextnode*/
ptr=ptr>next;
}
returncount;
}
Example 2: Write a program that travels a linked list consisting of
nodes of following struct type. Assume that the number of students in
the list is not known.

structstudent{
charname[15];
introll;
structstudent*next;
};
While travelling the list of students, it is split into two sub-lists pointed
by two pointers: front_half andback_half. The front_half points to the
front half of the list and the back_half to the back half of the list. If the
number of students is odd, the extra student should go into the front list.
Solution: The list of students would be created in the same fashion as
done in Example 1. The number of nodes would be counted during the
creation of the linked list itself. Assume that the linked list is being
pointed by First. The following steps would be followed to split the list in
the desired manner:
1. Computethemiddleofthelist.
2. Pointfront_halftoFirst.
3. PointanotherpointerFartoFirst.LetFartravelthe
liststartingfromthefirstandreachtothemiddle.
4. Pointback_halftothenodethatsucceedsthemiddleof
thelist.
5. AttachNULLtothenextpointerofthemiddlenode,now
pointedbyFar.
6. Printthetwolists,i.e.,pointedbyfront_halfand
back_half.

The required program is given below:

/*Thisprogramtravelsalinkedlistandsplitsit
intotwohalves*/
#include<stdio.h>
#include<alloc.h>
structstudent{
charname[15];
introll;
structstudent*next;
};
structstudent*create(int*count);
structstudent*split(structstudent*First,int
*count);

voiddispList(structstudent*First);
voidmain()
{structstudent*First,*front_half,*back_half;
intcount;
First=create(&count);
if(count==1)
{
printf(\nThelistcannotbesplit);
front_half=First;
}
else
{
front_half=First;
back_half=split(First,&count);
}
printf(\nTheFirstHalf&);
dispList(front_half);
printf(\nTheSecondHalf&);
dispList(back_half);
}
structstudent*create(int*count)
{
structstudent*First,*Far,*back;
intSz;
Sz=sizeof(structstudent);
/*Takefirstnode*/
First=(structstudent*)malloc(Sz);
/*Readthedataofthefirststudent*/
printf(\nEnterthedataofthefirststudent);
printf(\nName:);fflush(stdin);gets(First
>name);
printf(\nRoll:);scanf(%d,&First>roll);
First>next=NULL;
if(First>roll<=0){
printf(\nNodes=0);
First=NULL;
exit(1);}/*Emptylist*/
*count=1;
/*pointbackwhereFirstpoints*/
back=Far=First;

while(Far>roll>0)
{/*BringanewnodeinFar*/
Far=(structstudent*)malloc(Sz);
/*ReadDataofnextstudent*/
printf(\nName:);fflush(stdin);gets(Far
>name);
printf(\nRoll:);scanf(%d,&Far>roll);
if(Far>roll<=0)break;/*Endofthelist*/
*count=*count+1;
/*ConnectbacktoFar*/
back>next=Far;
/*pointbackwhereFarpoints*/
back=Far;
}/*repeattheprocess*/
back>next=NULL;
returnFirst;
}
structstudent*split(structstudent*First,int
*count)
{
structstudent*Far,*back;
intmiddle,i;
Far=First;/*Pointtothefirstnode*/
/*Splitthelist*/
{
/*computethemiddle*/
if((*count%2)==0)middle=*count/2;
elsemiddle=*count/2+1;
/*Travelthelistandsplitit*/
for(i=1;i<=middle;i++)
{
back=Far;
Far=Far>next;
}
back>next=NULL;
returnFar;
}
}
/*Printthelist*/
voiddispList(structstudent*First)

{
structstudent*Far;
Far=First;
while(Far!=NULL)
{
printf(%s,%d,Far>name,Far>roll);
Far=Far>next;
}
}
6.3.3 Searching a Linked List
Search is an operation in which an item is searched in the linked
list. In fact, the list is travelled from the first node to the last node. If
a visited node contains the item, then the search is successful and the
travel immediately stops. This means that in the worst case, whole of the
list would be travelled. An algorithm for searching an item X in the list
is given below.
In this algorithm a linked list, pointed by First, is travelled. While
travelling, the data part of each visited node is compared with the item
X. If the item is found, then the search stops otherwise the process
continues till the end of the list (i.e., NULL) is encountered. A
pointer ptr is used to visit the various nodes of the list.
Algorithm searchList()
{
Step
1.ifFirst==NULL{
printListEmpty;exit();}
2.ReadtheitemX
3.ptr=First;
4.flag=0;
5.while(ptr!=NULL)
{
5.1if(X==DATA(ptr)then{flag=1;break;}
5.2.ptr=NEXT(ptr);
}
6.if(flag==1)thenprintItemfoundelseprint
Itemnotfound
7.Stop
}

Example 3: Modify program of Example1 such that the program

searches the record of a student whose roll number is given by the user.
Solution: Instead of writing the code for creation and display
of linkedlists again and again, now onwards we would use the
functions create() and dispList() developed in earlier programs. A
new function based on algorithm searchList() would be written and
used to search the record of a student.
The required program is given below:

/*Thisprogramsearchesaniteminalinkedlist*/
#include<stdio.h>
#include<alloc.h>
#include<process.h>
structstudent{
charname[15];
introll;
structstudent*next;
};
structstudent*create();
intsearchList(structstudent*First,intRoll);
voidmain()
{
structstudent*First;
intRoll,result;
First=create();
printf(\nEntertheRollofthestudenttobe
searched);
scanf(%d,&Roll);
result=searchList(First,Roll);
if(result==1)
printf(\nTheStudentfound);
else
printf(\nTheStudentnotfound);
}
structstudent*create()/*Anyofthecreate()
functiondesignedin
aboveexamplescanbeused*/
{

}
intsearchList(structstudent*First,intRoll)
{structstudent*Far;
intflag;
Far=First;/*Pointtothefirstnode*/
flag=0;
while(Far!=NULL)
{
if(Roll==Far>roll)
{
flag=1;
break;
}
/*Pointtothenextnode*/
Far=Far>next;
}
if(flag==1)return1;
elsereturn0;
}
6.3.4 Insertion in a Linked List
We have already appreciated that linkedlists are most suitable and
efficient data structures for insertion and deletion operations. The
insertion of a node in a linked list involves the following three steps:
1. Take a node. Store data into the node.
2. Search the location in the list where the node is to be inserted.
3. Insert the node.

The location where the node is to be inserted in a linked list can either
be at the beginning or at any other arbitrary position in the list. An
algorithm that inserts a node at the beginning of the list is given below:
In this algorithm, a node pointed by ptr is inserted in a linked list
whose first node is pointed by the pointer First.
Algorithm insertAtHead()
{
Step
1.Takeanodeinptr.
2.ReadDATA(ptr).
3.If(First==NULL)then
{

3.1First=ptr;
3.2NEXT(First)=NULL;
}
else
{
3.1NEXT(Ptr)=First;
3.2First=ptr;
}
4.Stop
}
The list structure before and after the insertion operation is shown
in Figure 6.14.

Fig. 6.14 Insertion at the beginning of a list

It may be noted that after the insertion operation,
both First and ptr are pointing to first node of the list.
If it is desired to insert a node at arbitrary position in a linked list, then
it can either be inserted after anode or before a node in the list. An
algorithm for insertion of a node after a selected node in a linked list is
below.
Let First be the pointer to the linked list. In this algorithm, a node is
inserted after a node with data part equal to val. A pointer ptr
travels the list in such a way that each visited node is checked for data
part equal to val. If such a node is found then the new node, pointed by
a pointer called nptr, is inserted.
Algorithm insertAfter()
{
Step
1.ptr=First

2.while(ptr!=NULL)
{
2.1if(DATA(ptr)=val)
{takeanodeinnptr;
ReadDATA(nptr);
Next(nptr)=Next(ptr);
Next(ptr)=nptr;
break;
}
2.2ptr=Next(ptr);
}
3.Stop
}
The list structure before and after the insertion operation is shown
in Figure 6.15.

Fig. 6.15 Insertion after a selected node in the list

Example 4: Modify Example3 such that after creation of a linked list
of students, the program asks from the user to enter data for the student
which is desired to be inserted into the list after a givennode.
Solution: The required program is given below. It takes care of both the
cases, i.e., it inserts the student as per his roll number in the list, which
can be at the beginning or at any other location in the list.
/*Thisprograminsertsaniteminalinkedlist*/

#include<stdio.h>
#include<alloc.h>
#include<process.h>
structstudent{
charname[15];
introll;
structstudent*next;
};
structstudent*create();
structstudent*insertList(structstudent*First,
intRoll,int*flag);
voiddispList(structstudent*First);
voidmain()
{
structstudent*First;
intRoll,result;
First=create();
if(First!=NULL)/*Insertinbetween*/
{
printf(\nEntertherollno.ofthestudent);
printf(\nafterwhichtheinsertionistobe
made:);
scanf(%d,&Roll);
}
First=insertList(First,Roll,&result);
if(result==1)
{printf(\nThelist.);
dispList(First);
}
else
printf(\nTheplacenotfound);
}
structstudent*create()/*Hereanyofthe
create()functiondesigned
inaboveexamplescanbeused*/
{
}
structstudent*insertList(structstudent*First,int
Roll,int*flag)
{
structstudent*ptr,*Far;

intSz;/*InsertattheHeadofthelist*/
if((First==NULL)||(First>roll>Roll))
{
ptr=(structstudent*)malloc(Sz);
/*ReadDataofnextstudent*/
printf(\nName:);fflush(stdin);gets(ptr
>name);
printf(\nRoll:);scanf(%d,&ptr>roll);
ptr>next=First;
First=ptr;
*flag=1;/*indicatesthatinsertionwasmade*/
}
else
{Far=First;/*Pointtothefirstnode*/
*flag=0;
while(Far!=NULL)
{
if(Roll==Far>roll)
{
*flag=1;
break;/*Thelocationoftheinsertion
found*/
}
/*Pointtothenextnode*/
Far=Far>next;
}
if(*flag==1)
{
ptr=(structstudent*)malloc(Sz);
/*ReadDataofstudentforinsertion*/
printf(\nEnterthedataofstudent);
printf(\nName:);fflush(stdin);gets(ptr
>name);
printf(\nRoll:);scanf(%d,&ptr>roll);
ptr>next=Far>next;/*Insertnode*/
Far>next=ptr;
}
}
ptr=First;
returnptr;
}

voiddispList(structstudent*First)/*Hereanyof
thedispList()
functiondesignedinaboveexamplescanbeused*/
{
}
Note: It is easier to insert a node after a selected node in a linked list as
only one pointer is needed to select the node. For example, in the above
program, the Far pointer has been used to select the nodeafter which
the insertion is made.
However, if insertion is needed to be made before a selected node then it
is cumbersome to insertnode with only one pointer. The reason being
that when the node is found before which the insertion is desired, there
is no way to reach to its previous node and make the insertion. The
easier and elegant solution is to use two pointers (Far and back) which
move in tandem, i.e., the pointer Far is followed by back.
When Far reaches the desired location, back points to its previous
location as shown inFigure 6.16.

Fig. 6.16 Insertion before a given node in a linked list

An algorithm for insertion of data before a selected node in a linked list
pointed by First is given below:
Algorithm insertBefore()
{
Far=First;
If(DATA(Far)==Val)/*Checkifthefirstnodeis
thedesirednode*/
{
takeanewnodeinptr;
ReadDATA(ptr);
Next(ptr)=Far;
First=ptr;
Stop;
}
while(Far!=NULL)
{

back=Far;
Far=Next(Far);
If(DATA(Far)==Val)/*Checkifthenodeisthe
desirednode*/
{
takeanewnodeinptr;
ReadDATA(ptr);
Next(ptr)=Far;
Next(back)=ptr;
break;
}
}
}
Example 5: A sorted linked list of students in the order of their roll
numbers is given. Write a program that asks from the user to
enter data for a missing student which is desired to be inserted into the
list at its proper place.
Solution: The sorted linked list of students in the order of their roll
numbers would be created by using function create(), developed in
the previous programs. The data of the student which is to be inserted is
read and its proper location found, i.e., before the node whose roll
number is greater than the roll number of this in coming student. The
algorithm insertBefore() would be used for the desired insertion
operation.
The required program is given below:
/*Thisprograminsertsanodeinalinkedlist*/
#include<stdio.h>
#include<alloc.h>
#include<process.h>
structstudent{
charname[15];
introll;
structstudent*next;
};
structstudent*create();
structstudent*insertList(structstudent*First,
structstudentnewStud,
int*flag);
voiddispList(structstudent*First);

voidmain()
{
structstudent*First,stud;
intRoll,result;
First=create();
printf(\nEnterthedataofthestudentwhichisto
beinserted);
scanf(%s%d,&stud.name,&stud.roll);
First=insertList(First,stud,&result);
if(result==1)
{printf(\nThelist.);
dispList(First);
}
else
printf(\nTheplacenotfound);
}
structstudent*create()/*Hereanyofthe
create()functiondesigned
inaboveexamplescanbeused*/
{
}
structstudent*insertList(structstudent*First,
structstudentnewStud,
int*flag)
{
structstudent*ptr,*Far,*back;
intSz;/*InsertattheHeadofthelist*/
if((First==NULL)||(First>roll>newStud.roll))
{
Far=(structstudent*)malloc(Sz);
/*ReadDataofnextstudent*/
strcpy(Far>name,newStud.name);
Far>roll=newStud.roll;
Far>next=First;
First=Far;
*flag=1;/*indicatesthatinsertionwasmade*/
}
else
{ptr=back=First;/*Pointtothefirstnode*/
*flag=0;
while(ptr!=NULL)

{
ptr=ptr>next;
if(newStud.roll>back>roll&&newStud.roll<ptr
>roll)
{
*flag=1;
break;/*Thelocationoftheinsertionfound*/
}
back=ptr;
}
if(*flag==1)
{
Far=(structstudent*)malloc(Sz);
/*ReadDataofstudentforinsertion*/
strcpy(Far>name,newStud.name);
Far>roll=newStud.roll;
Far>next=ptr;/*Insertnode*/
back>next=Far;
}
}
ptr=First;
returnptr;
}
voiddispList(structstudent*First)/*Hereanyof
thedispList()function
designedinaboveexamplescanbeused*/
{
}
6.3.5 Deleting a Node from a Linked List
A delete operation involves the following two steps:
1. Search the list for the node which is to be deleted.
2. Delete the node.

The first step requires that two pointers in tandem be used to search
the node needs to be deleted. After the search, one pointer points to the
selected node and the other points to immediate previousnode.
An algorithm for deletion of a node from a list pointed by First is given
below. It uses two pointers,ptr and back, that travel the list in such a
way that each visited node is checked. If it is the node to be deleted,

then ptr points to this selected node and the pointer back points to its
immediate previousnode as shown in Figure 6.17.
Algorithm delNode()
{
1.if(DATA(First)=VAL/*Checkifthestarting
nodeisthedesired
one*/
{
1.1ptr=First;
1.2First=Next(First);
1.3Free(ptr);
1.4Stop;
}
2.Back=First;
3.ptr=Next(First)
4.while(ptr!=NULL)
{
4.1if(DATA(ptr)=VAL)
{
Next(Back)=Next(ptr);/*Thenodeisdeleted*/
Free(ptr);
break;
}
4.2back=ptr;
4.3ptr=Next(ptr);
}
}
5.Stop
}
The list structure before and after the deletion operation is shown
in Figure 6.17.

Fig. 6.17 The deletion operation

Example 6: A linked list of students in the order of their roll numbers is
given. Write a program that asks from the user to enter roll number of
the student which is desired to be deleted from the list.
Solution: The linked list of students in the order of their roll numbers
would be created by using function create(), developed in the
previous programs. The roll number of the student which is to be deleted
is read and its proper location found. The algorithm delNode() would
be used for the desired deletion operation.
The required program is given below:
/*Thisprogramdeletesanodefromalinkedlist*/
#include<stdio.h>
#include<alloc.h>
#include<process.h>
structstudent{
charname[15];
introll;
structstudent*next;
};
structstudent*create();
structstudent*delNode(structstudent*First,int
roll,int*flag);
voiddispList(structstudent*First);

voidmain()
{
structstudent*First;
intRoll,result;
First=create();
printf(\nEntertheRollofthestudentwhichisto
bedeleted);
scanf(%d,&Roll);
First=delNode(First,Roll,&result);
if(result==1)
{printf(\nThelist.);
dispList(First);
}
else
printf(\nTheplacenotfound);
}
structstudent*create()/*Hereanyofthecreate()
functiondesigned
inaboveexamplescanbeused*/
{
}
structstudent*delNode(structstudent*First,int
Roll,int*flag)
{
structstudent*ptr,*Far,*back;
intSz;
if(First==NULL){*flag=0;returnFirst;}
/*deletefromtheHeadofthelist*/
if(First>roll==Roll)
{
ptr=First;
First=First>next;
free(ptr);
*flag=1;/*indicatesthatdeletionwasmade*/
}
else
{ptr=back=First;/*Pointtothefirstnode*/
*flag=0;
while(ptr!=NULL)
{

ptr=ptr>next;
if(Roll==ptr>roll)
{
*flag=1;
break;/*Thelocationofthedeletionfound*/
}
back=ptr;
}
if(*flag==1)
{back>next=ptr>next;
free(ptr);
}
}
returnFirst;
}
voiddispList(structstudent*First)/*Hereanyof
thedispList()function
designedinaboveexamplescanbeused*/
{
}
6.4 VARIATIONS OF LINKED LISTS

There are many limitations of a linear linked list. Some of them are as
follows:
1. The movement through the linked list is possible only in one direction, i.e.,
in the direction of the links.
2. As discussed above, a pointer is needed to be pointed to a node before
the node that needs to be deleted.
3. Insertion or appending of a node at the end of the linked list can only be
done after travelling whole of the list.

In order to overcome the limitations, many variations of linear or

singly linkedlists have been suggested in the literature. Some of the
popular variations of linear linkedlists are given in subsequent
sections.
6.4.1 Circular Linked Lists
A circular linked list is a linked list in which the last node points to
the node at the beginning of the list as shown in Figure 6.18. Thus, it

forms a circle, i.e., there is no NULL pointer. In fact, every node has a
successor.

Fig. 6.18 Circular linked list

The main advantage of the circular linked list is that from any node, in
the list, one can reach to any other node. This was not possible in a
linear linked list, i.e., there was no way of reaching to nodes preceding
the current node.
A modified algorithm for creation of circular linked list is given below:
Algorithm createCirLinkList()
{
Step
1.Takeanewnodeinpointercalledfirst.
2.ReadData(First).
3.Pointbackpointertothesamenodebeing
pointedbyfirst,i.e.,back=first.
4.Bringanewnodeinthepointercalledfar.
5.ReadData(far).
6.Connectnextofbacktofar,i.e.,back>next
=far.
7.Takebacktofar,i.e.,back=far.
8.Repeatsteps4to7tillwholeofthelistis
constructed.
9.PointnextoffartoFirst,i.e.,far>next=
First.
10.Stop.
}
Example 7: Write a function create() that creates a circular linked
list of N students of following structure type:
structstudent{
charname[15];
introll;
structstudent*next;
};
Solution: We have used the algorithm createCirLinkList() to
write the required function which is given below:

structstudent*create()
{
structstudent*First,*Far,*back;
intN,i;
intSz;
Sz=sizeof(structstudent);
printf(\nEnterthenumberofstudentsinthe
class);
scanf(%d,&N);
if(N==0)returnNULL;/*Listisempty*/
/*Takefirstnode*/
First=(structstudent*)malloc(Sz);
/*Readthedataofthefirststudent*/
printf(\nEnterthedata);
printf(\nName:&#x201D;);fflush(stdin);
gets(First>name);
printf(\nRoll:&#x201D;);scanf(%d,&First
>roll);
First>next=NULL;
/*pointbackwhereFirstpoints*/
back=First;
for(i=2;i<=N;i++)
{/*BringanewnodeinFar*/
Far=(structstudent*)malloc(Sz);
/*ReadDataofnextstudent*/
printf(\nName:&#x201D;);fflush(stdin);
gets(Far>name);
printf(\nRoll:&#x201D;);scanf(%d,&Far
>roll);
/*ConnectBacktoFar*/
back>next=Far;
/*pointbackwhereFarpoints*/
back=Far;
}/*Repeattheprocess*/
Far>next=First;/*Pointthelastpointertothe
firstnode*/
returnFirst;/*returnthepointertothelinked
list*/
}

Example 8: Write a program that travels a circular linked list

consisting of nodes of following structtype. While travelling, it counts
the number of nodes in the list. Finally, the count is printed.

structstudent{
charname[15];
introll;
structstudent*next;
};
Solution: A circular linked list of nodes of student type would be
created by using the function calledcreate() developed in Example 7.
The list would be travelled from the first node till we reach backto the
first node. While travelling, the visited nodes would be counted. The
required program is given below:
/*Thisprogramtravelsacircularlinkedlistand
countsthenumberof
nodes*/
#include<stdio.h>
#include<alloc.h>
structstudent{
charname[15];
introll;
structstudent*next;
};
structstudent*create();
intcountNode(structstudent*First);
voidmain()
{
intcount;
structstudent*First;
First=create();
count=countNode(First);
printf(\nThenumberofnodes=%d,count);
}
structstudent*create()/*Herethecreate()
functionofExample7can
beused*/
{

}
/*Travelandcountthenumberofnodes*/
intcountNode(structstudent*First)
{
structstudent*ptr;
intcount;
count=0;
ptr=First;/*Pointtothefirstnode*/
do
{count=count+1;
/*Pointtothenextnode*/
ptr=ptr>next;
}
while(ptr!=First);/*Thetravelstopsatthe
firstnode*/
returncount;
}
Note: The above representation of circular linkedlist (Figure 6.18)
has a drawback as far as the insertion operation is concerned. Even if
the node is to be added at the head of the list, the complete list will have
to be travelled so that the last pointer is made to point to the new node.
This operation ensures that the list remains circular after the insertion.
It may be noted that this extra travel was not required in the case of
linear linked list.
The above drawback of circular linkedlists can be avoided by
shifting the First pointer to the lastnode in the list as shown in Figure
6.19.
It may be noted that by having this representation, both ends of the list
can be accessed through theFirst. The First points to one end and
the Next(First) points to another.

Fig. 6.19 Another representation of a circular linked list

A node can be added at the head of the list by the following algorithm:
Algorithm insertHead()
{
Step

1.Takeanewnodeinptr.
2.ReadDATA(ptr).
3.Next(ptr)=Next(First).
4.Next(First)=ptr.
}
The effect of the above algorithm is shown in Figure 6.20.

Fig. 6.20 Insertion of a node at the head of a circular linked list without
travel
A node can be added at the tail (i.e., last) of the list by the following
algorithm:
Algorithm insertTail()
{
Step
1.Takeanewnodeinptr.
2.ReadDATA(ptr).
3.Next(ptr)=Next(First).
4.Next(First)=ptr.
5.First=ptr.
}
The effect of the above algorithm is shown in Figure 6.21.

Fig. 6.21 Insertion of a node at the tail of a circular linked list without
travel
It may be noted that in both the cases of insertion (at the head and tail),
no travel of the circular linked list was made. Therefore, the new
representation has removed the drawback of the ordinary circular linked
list. In fact, it offers the solution in the order of O(1).
6.4.2 Doubly Linked List
Another variation of linked list is a doubly linked list. It is a list in which
each node contains two links:leftLink and rightLink.
The leftLink of a node points to its preceding node and
therightLink points to the next node in the list (see Figure 6.22).
Fig. 6.22 Node of a doubly linked list
From Figure 6.22, it can be noticed that a node in the doubly linked list
has the following three fields:
1. Data: for the storage of information.
2. leftLink:
pointer to the preceding node.
3. rightLink:
pointer to the next node.

The arrangement of nodes in a doubly linked list is shown in Figure 6.23.

Fig. 6.23 A doubly linked list

The algorithm for creation of doubly linked list is given below:
Algorithm createDList()
{
Step
1.TakeanewnodeinpointercalledFirst.
2.PointleftLinkoffirsttoNULL,i.e.,first
>leftLink=NULL.
3.ReadData(First).
4.Pointbackpointertothenodebeingpointedby
First,i.e.,back=First.
5.BringanewnodeinthepointercalledFar.
6.ReadData(Far).
7.ConnectrightLinkofbacktoFar,i.e.,back
>rightLink=Far.
8.ConnectleftLinkofFartoBack,i.e.,Far
>leftLink=Back.
9.TakebacktoFar,i.e.,back=Far.
10.Repeatsteps5to9tillwholeofthelistis
constructed.
11.PointrightLinkoffartoNULL,i.e.,Far>
rightLink=NULL.
12.Stop.
}
Example 9: Write a program that uses a function create() to create
a doubly linked list of Nstudents of following structure type.
structstudent{
charname[15];
introll;
structstudent*leftList,*rightList;
};
It travels the list in both directions, i.e., forward and backward. The data
part of visited nodes is displayed.

Solution: We have used the algorithm createDList() to write the

required program. Two functions
dispForward() and dispBackward()are also written to
travel the list in both the directions. The required program is given
below:
/*Thisprogramcreatesadoublylinkedlistof
students*/
#include<stdio.h>
#include<alloc.h>
structstudent{
charname[15];
introll;
structstudent*leftList,*rightList;
};
structstudent*create();
voiddispForward(structstudent*First);
voiddispBackward(structstudent*First);
voidmain()
{
structstudent*First;
First=create();
dispForward(First);
dispBackward(First);
}
structstudent*create()
{
structstudent*First,*Far,*back;
intN,i;
intSz;
Sz=sizeof(structstudent);
printf(\nEnterthenumberofstudentsinthe
class);
scanf(%d,&N);
if(N==0)returnNULL;/*Listisempty*/
/*Takefirstnode*/
First=(structstudent*)malloc(Sz);
/*Readthedataofthefirststudent*/
printf(\nEnterthedata);
printf(\nName:);fflush(stdin);gets(First
>name);

printf(\nRoll:);scanf(%d,&First>roll);
First>rightList=NULL;
First>leftList=NULL;
/*pointbackwhereFirstpoints*/
back=First;
for(i=2;i<=N;i++)
{/*BringanewnodeinFar*/
Far=(structstudent*)malloc(Sz);
/*ReadDataofnextstudent*/
printf(\nName:);fflush(stdin);gets(Far
>name);
printf(\nRoll:);scanf(%d,&Far>roll);
/*ConnectBacktoFar*/
back>rightList=Far;
/*pointbackwhereFarpoints*/
Far>leftList=back;
back=Far;
}/*Repeattheprocess*/
Far>rightList=NULL;
returnFirst;
}
/*Printthecreatedlinkedinforwarddirection*/
voiddispForward(structstudent*First)
{structstudent*Far;
Far=First;/*Pointtothefirstnode*/
printf(\nTheForwardListis.);
while(Far!=NULL)
{
printf(%s%d,,Far>name,Far>roll);
/*Pointtothenextnode*/
Far=Far>rightList;
}
}
/*Printthecreatedlinkedinbackward
direction*/
voiddispBackward(structstudent*First)
{structstudent*Far;
Far=First;/*Pointtothefirstnode*/
while(Far>rightList!=NULL)
{/*Traveltothelastnode*/
Far=Far>rightList;

}
printf(\nTheBackwordListis.);
while(Far!=NULL)/*Travelthelistinbackward
direction*/
{
printf(%s%d,,Far>name,Far>roll);
/*Pointtothenextnode*/
Far=Far>leftList;
}
}
Consider the doubly linked list given in Figure 6.24.
The following relations hold good for the pointer ptr:
1. ptr ptr -> leftLink -> rightLink
2. ptr ptr -> rightLink -> leftLink

The first relation indicates that the rightLink of leftLink of ptr is

pointing to the same nodebeing pointed by ptr. Similarly, the second
relation indicates that the leftLink of rightLink ofptr is pointing to
the same node being pointed by ptr.
The above relations help us to insert node in a doubly linked list. The
discussion on this aspect is given in Section 6.4.2.1.

Fig. 6.24 The left of right and right of left is the same node
6.4.2.1 Insertion in Doubly Linked List
Insertion of a node X before a node pointed by ptr can be done by the
following program segment:
1.
2.
3.
4.

X -> rightLink = ptr -> leftLink -> rightLink;

ptr -> leftLink -> rightLink = X;
X -> leftLink = ptr -> leftLink;
ptr -> leftLink =X;

The effect of the above program segment is shown in Figure 6.25. The
above four operations have been labelled as 14 in Figure 6.25.

Fig. 6.25 Insertion of node X before a node pointed by ptr

Similarly, insertion of a node X after a node pointed by ptr can be done
by the following program segment:
1.
2.
3.
4.

X -> leftLink = ptr ->rightLink ->leftLink;

ptr ->rightLink ->leftLink = X;x
X ->rightLink = ptr ->rightLink;
ptr ->rightLink =X;

The effect of the above program segment is shown in Figure 6.26. The
above four operations have been labeled as 14 in Figure 6.26.
6.4.2.2 Deletion in Doubly Linked List
Deletion of a node pointed by ptr can be done by the following program
segment:
1. ptr>left>right=ptr>right;
2. ptr>right>left=ptr>left;
3. free(ptr);

Fig. 6.26 Insertion of node X after a node pointed by ptr

The effect of the above program segment is shown in Figure 6.27. The
above three operations have been labeled as 13 in Figure 6.27.

Fig. 6.27 Deletion of a node pointed by ptr

Note: During insertion or deletion of a node,
the leftLink and rightLink pointers are appropriately fixed so that
they point to the correct node in the list.
From Figures 6.23 to 6.27, it may be observed that the doubly linked list
has the following advantages over the singly linked list:
1.
2.
3.
4.

With the help of rightLink, the list can be traversed in forward direction.
With the help of leftLink, the list can be traversed in backward direction.
Insertion after or before a node is easy as compared to singly linked list.
Deletion of a node is also very easy as compared to singly linked list.

It is left as an exercise for the readers to implement the insertion and

deletion operations using C.
6.5 THE CONCEPT OF DUMMY NODES

The linked list and its variants, discussed above, have a major drawback.
Every algorithm has to take care of special cases. For instance, besides a
normal operation, an algorithm has to make sure that it works for the
following exceptional or boundary cases:
1. the empty list
2. on the head element of the list
3. on the last element of the list

The above situation occurs when a node in a linked list has no

predecessor or a successor. For example, the first node in a linked list
has no predecessor and, therefore, insertion or deletion at the end
requires a special case. Similarly, an empty circular linked list will not be
circular. Same is the case with doubly linked list.

The above anomalous situation can be resolved with the help of

a dummy node. The dummy node is a head or front node which is
purposefully left blank. The First pointer always points to the
dummynode as shown in Figure 6.28.

Fig. 6.28 Linked list with a dummy node

It may be noted that the list shown in Figure 6.28 represents a list of
integers (25, 37, 68 ). The dummy node does not contain any value. It
may be further noted that first node of the list (containing 25) is also
having dummy node as its predecessor and, therefore, insertion at the
head of the node can be done in a normal fashion as done in
algorithm insertAfter() of section 6.3.4.
Similarly, dummy nodes can also be used for circular and
doubly linkedlists as shown in Figure 6.29.

Fig. 6.29 Circular and doubly linkedlists with dummy nodes

Note: With the inclusion of a dummy node, an algorithm does not have
to check for special or exceptional cases such as existence of empty list or
insertion/deletion at the head of the list and the like. Thus, the algorithm
becomes easy and applicable to all situations.
Some programmers employ the dummy node to store useful information
such as number of elements in the list or the name of the group or set to
which the elements belong to. For example, the list shown in Figure
6.28 can be modified so that the dummy node contains information
about number of elements present in the list (3 in this case). The
modified list is given in Figure 6.30.

Fig. 6.30 Dummy node contains important information

Many programmers use a dummy node called a trailer node at the tail
end so that each node in the list has a successor.
Another benefit of dummy node is that an empty circular linked list is
also circular in nature. Same is the case with doubly linked list as shown
in Figure 6.31.

Fig. 6.31 Comparison of empty lists with and without dummy nodes
It may be noted that with the help of dummy nodes, the basic nature
of linkedlists remains intact.
6.6 LINKED STACKS

We know that a stack is a LIFO (last in first out) data structure. Earlier in
the book, the stack was implemented using an array. Though the array
implementation was efficient but a considerable amount of storage space
was wasted in the form of unutilized locations of the array. For example,
let us assume that at a given time, there are five elements in a stack. If
the stack is implemented using 20 locations, then 15 locations out of 20
are being wasted.

Fig. 6.32 A linked stack

A solution to this problem is dynamic storage allocation, i.e., implement
stack using linkedlists. Since a stack involves frequent additions
and deletions at one end called Top, the linked list is definitely a better
choice from the point of implementation.
Obviously, the structure of the node in a linked stack will also be selfreferential. The arrangement of nodes is shown in Figure 6.32. It may be
noted that its one end is being pointed by a pointer called Topand the
other end is pointing to NULL.
An algorithm for pushing an element on to the stack is given below. In
this algorithm, a node pointed by ptr is pushed onto the stack.
Algorithm Push()
{
Step
1.Takeanewnodeinptr;
2.ReadDATA(ptr);
3.Next(ptr)=Top;
4.Top=ptr;
}
The structure of linked stack after a push operation is shown in Figure
6.33.
An algorithm for popping an element from the stack is given below. In
this algorithm, a node from the Top of the stack is removed. The
removed node is pointed by ptr.
Algorithm POP()
{
Step

1.if(Top=NULL)thenpromptstackempty;Stop
2.ptr=Top
3.Top=Next(Top);
4.returnDATA(ptr)
5.Freeptr
}
The structure of linked stack after a POP operation is shown in Figure
6.34.

Fig. 6.33 Linked stack after a PUSH operation

Fig. 6.34 Linked stack after a POP operation

Example 10: Write a menu driven program that simulates a linked
stack for integer data items.
Solution: The algorithms Push() and POP() developed in this
section would be used to write the program that simulates a linked stack
for integer data items. Each node of the linked stack would be of
following structure type:

structnode
{
intitem;
structnode*next;
};
Therequiredprogramisgivenbelow:
/*Thisprogramsimulatesalinkedstackforinteger
dataitems*/
#include<stdio.h>
#include<alloc.h>
#include<conio.h>
structnode
{
intitem;
structnode*next;
};
structnode*push(structnode*Top,intitem);
structnode*pop(structnode*Top,int*item);
voiddispStack(structnode*Top);
voidmain()
{
intitem,choice;
structnode*Top;
Top=NULL;
do
{clrscr();
printf(\nMenu);
printf(\nPush1);
printf(\nPop2);
printf(\nDisplay3);
printf(\nQuit4);
printf(\n\EnterChoice:);
scanf(%d,&choice);
switch(choice)
{
case1:printf(\nEntertheintegeritemtobe
pushed:);

scanf(%d,&item);
Top=push(Top,item);
break;
case2:if(Top==NULL)
{printf(\nStackempty);
break;
}
Top=pop(Top,&item);
printf(\nThepoppeditem=%d,item);
break;
case3:dispStack(Top);
break;
}
printf(\npressanykey);getch();
}
while(choice!=4);
}
structnode*push(structnode*Top,intitem)
{
structnode*ptr;
intsize;
size=sizeof(structnode);
ptr=(structnode*)malloc(size);
ptr>item=item;
ptr>next=Top;
Top=ptr;
returnTop;
}
structnode*pop(structnode*Top,int*item)
{
structnode*ptr;
if(Top==NULL)returnNULL;
ptr=Top;
Top=Top>next;
*item=ptr>item;
free(ptr);
returnTop;
}
voiddispStack(structnode*Top)
{
if(Top==NULL)

printf(\nStackEmpty);
else
{
printf(\nTheStackis:);
while(Top!=NULL)
{
printf(%d,Top>item);
Top=Top>next;
}
}
}
6.7 LINKED QUEUES

We know that a queue is a FIFO (first in first out) data structure. Since
this data structure also involves deletions and additions at front end
and rear end respectively, a linked list is definitely a better choice from
the point of implementation.
A linked queue will have the same node structure as in the case of a
stack as shown in Figure 6.35. Two pointers called Front and Rear would
keep track of the front and rear end of the queue.

Fig. 6.35 A linked queue

An algorithm for adding an element on to the queue is given below. In
this algorithm, a node pointed by ptr is added into the queue.
Algorithm AddQueue()
{
Step
1.Takeanewnodeinptr;
2.ReadDATA(ptr);
3.Next(ptr)=Next(Rear);
4.Next(Rear)=ptr;
5.Rear=ptr;
}
The structure of linked queue after addition operation is shown in Figure
6.36.

Fig. 6.36 The linked queue after an addition operation

An algorithm for deleting an element from the linked queue is given
below. In this algorithm, a nodefrom the front of the queue is removed.
Algorithm delQueue()
{
Step
1.if(Front==Rear)thenpromptQueueempty;
Stop
2.ptr=Front;
3.Front=Next(Front);
4.Freeptr
5.returnDATA(Front)
}
The structure of linked queue after a deletion operation is shown
in Figure 6.37.

Fig. 6.37 The linked queue after a deletion operation

It may be noted that the Front always points to a blank node, i.e., which
does not contain any data. Thus, when the queue is empty then both
Front and Rear point to a blank node.
Example 11: Write a menu-driven program that simulates a linked
queue for integer data items.
Solution: The algorithms AddQueue() and delQueue() developed
in this section would be used to write the program that simulates a
linked queue for integer data items. Each node of the linked queue
would be of following structure type:
structnode
{
intitem;
structnode*next;
};
Therequiredprogramisgivenbelow:
/*Thisprogramsimulatesalinkedqueueforinteger
dataitems*/
#include<stdio.h>
#include<alloc.h>
#include<conio.h>

structnode
{
intitem;
structnode*next;
};
structnode*addQ(structnode*Rear,structnode
*Front,intitem);
structnode*delQ(structnode*Rear,structnode
*Front,int*item);
voiddispQ(structnode*Rear,structnode*Front);
voidmain()
{
intitem,choice,size;
structnode*Front,*Rear;
size=sizeof(structnode);
Front=(structnode*)malloc(size);
Front>item=9999;/*dummynode*/
Front>next=NULL;
Rear=Front;/*InitializeQueuetoempty*/
do
{clrscr();
printf(\nMenu);
printf(\nAdd1);
printf(\nDelete2);
printf(\nDisplay3);
printf(\nQuit4);
printf(\n\nEnterChoice:);
scanf(%d,&choice);
switch(choice)
{
case1:printf(\nEntertheintegeritemtobe
added:);
scanf(%d,&item);
Rear=addQ(Rear,Front,item);
break;
case2:if(Front==Rear)
{printf(\nQueueempty);
break;
}

Front=delQ(Rear,Front,&item);
printf(\nThedeleteditem=%d,item);
break;
case3:dispQ(Rear,Front);
break;
}
printf(\npressanykey);getch();
}
while(choice!=4);
}
structnode*addQ(structnode*Rear,structnode
*Front,intitem)
{
intsize;
structnode*ptr;
size=sizeof(structnode);
ptr=(structnode*)malloc(size);
ptr>item=item;
ptr>next=Rear>next;
Rear>next=ptr;
Rear=ptr;
returnptr;
}
structnode*delQ(structnode*Rear,structnode
*Front,int*item)
{
structnode*ptr;
ptr=Front;
Front=Front>next;
*item=Front>item;
Front>item=9999;/*Thedummynode*/
free(ptr);
returnFront;
}
voiddispQ(structnode*Rear,structnode*Front)
{
if(Front==Rear)
printf(\nThequeueempty);
else
{printf(\nThequeueis...);
do

{
Front=Front>next;
printf(%d,Front>item);
}
while(Front!=Rear);
}
}
6.8 COMPARISON OF SEQUENTIAL AND LINKED STORAGE

A list can be stored in a one-dimensional array and in a linked list as

well. The choice of data structure depends upon the problem at hand and
the type of operations to be performed on the data. A summarized
comparison of sequential and linked storage is given below:
1. Unlike arrays, the linked list is a dynamic data structure. The linked list
implementation adds the elements to the list by getting the memory
allocated dynamically. Therefore, the number of elements present in the
linked list is limited only by the memory available to the operating system.
We can say that the limitation of arrays has been effectively handled by the
linked list implementation.
2. linkedlists store the elements in non-contiguous locations. The nodes
are independent of each other.
3. linkedlists are useful for storing elements in the order of insertion and
subsequent traversal in that order.
4. However, the linkedlists occupy more space than the arrays because
each node contains an extra pointer besides the information.
5. Compared to arrays, linkedlists are slow in many ways. The dynamic
memory is allocated through a system call to operating system making the
processing very slow. Moreover, every element is referred to through an
extra reference made through a pointer thereby making the processing of
linked list further slow.
6. Access to an element in a linked list is purely sequential whereas an element
in array can be accessed through its index. Thus, arrays are faster in
accessing an element.
7. Nevertheless, a linked list offers a very easy way of insertions and deletions
in a list without the physical shifting of elements in the list.
8. Without moving the nodes, the elements of the list can be reordered by
simply changing the pointers.

6.9 SOLVED PROBLEMS

Problem1: Write an algorithm which travels a linked list pointed

by List. It travels the list in such a way that it reverses the links of the
visited nodes, i.e., at the end the list becomes reversed.
Solution: An algorithm called revLink( ) is given below. In this
algorithm, a linked list pointed by the pointer List is travelled with the
help of three pointers back, Ahead, and ptr. During the travel, the
direction of link of each visited node is reversed, i.e., it points to its
immediate preceding node. At the end, the pointer List is made to point
to the last node as shown in Figure 6.38.
Algorithm revLink()
{
Step
1.if(List==NULL)returnList;
2.ptr=List;
3.back=List;
4.Ahead=next(ptr);
5.next(List)=NULL;
6.while(Ahead!=NULL)repeatsteps7to10
7.ptr=Ahead;
8.Ahead=next(Ahead)
9.next(ptr)=back;/*reversethelink*/
10.back=ptr;
11.List=ptr;
12.Stop;
}

Fig. 6.38 Reversal of links of a linked list

Problem 2: Write a program which travels a linked list pointed

by List. It travels the list in such a way that it reverses the links of the
visited nodes, i.e., at the end the list becomes reversed and its contents
are displayed.
Solution: The algorithm revLink() developed in Problem1 is
used to write the given program that reverses a linked list of students
pointed by a pointer called List. The required function is given below:
/*Thisfunctionreversesthelinksofalinkedlist
ofstudents*/
#include<stdio.h>
#include<alloc.h>
structstudent{
charname[15];
introll;
structstudent*next;
};
structstudent*create();
structstudent*revLink(structstudent*List);
voiddispList(structstudent*First);
voidmain()
{
structstudent*First;
First=create();
First=revLink(First);
dispList(First);
}
structstudent*create()
{
structstudent*First,*Far,*back;
intN,i;
intSz;
Sz=sizeof(structstudent);
printf(\nEnterthenumberofstudentsinthe
class);
scanf(%d,&N);
if(N==0)returnNULL;/*Listisempty*/
/*Takefirstnode*/
First=(structstudent*)malloc(Sz);
/*Readthedataofthefirststudent*/
printf(\nEnterthedata);

printf(\nName:);fflush(stdin);gets(First
>name);
printf(\nRoll:);scanf(%d,&First>roll);
First>next=NULL;
/*pointbackwhereFirstpoints*/
back=First;
for(i=2;i<=N;i++)
{/*BringanewnodeinFar*/
Far=(structstudent*)malloc(Sz);
/*ReadDataofnextstudent*/
printf(\nName:);fflush(stdin);gets(Far
>name);
printf(\nRoll:);scanf(%d,&Far>roll);
/*ConnectBacktoFar*/
back>next=Far;
/*pointbackwhereFarpoints*/
back=Far;
}/*repeattheprocess*/
Far>next=NULL;
returnFirst;
}
structstudent*revLink(structstudent*List)/*The
functionthatreverses
thelist*/
{
structstudent*back,*ptr,*ahead;
if(List==NULL)
{
printf(\nThelistisempty);
returnNULL;
}
ptr=List;
back=List;
ahead=ptr>next;
List>next=NULL;
while(ahead!=NULL)
{
ptr=ahead;
ahead=ahead>next;
ptr>next=back;/*Connectthepointerto
precedingnode*/

back=ptr;
}
List=ptr;
returnList;
}
/*Printthecreatedlinkedlist*/
voiddispList(structstudent*First)
{structstudent*Far;
Far=First;/*Pointtothefirstnode*/
printf(\nTheListis.);
while(Far!=NULL)
{
printf(%s%d,,Far>name,Far>roll);
/*Pointtothenextnode*/
Far=Far>next;
}
}
Problem 3: Write a program that adds two polynomials Pol1and
Pol2, represented as linkedlists of terms, and gives a third
polynomial Pol3suchthatPol3=Pol1+Pol2.
Solution: We will use the following structure for the term of a
polynomial:
As done in Chapter 3, the polynomials will be added by merging the
two linkedlists. The required program is given below:
/*Thisprogramaddstwopolynomialsrepresentedas
linkedlistspointedbytwopointerspol1andpol2
andgivesathirdpolynomialpointedbyathird
pointerpol3*/
#include<stdio.h>
structterm
{
intcoef;
intexp;
structterm*next;
};
structterm*create_pol();
voidmain()

{
structterm*pol1,*pol2,*pol3,*ptr1,*ptr2,*ptr3,
*back,*bring,*ptr;
intsize;
size=sizeof(structterm);
printf(\nPolynomial1);
pol1=create_pol();
printf(\nPolynomial2);
pol2=create_pol();
ptr1=pol1;
ptr2=pol2;
ptr3=pol3=NULL;
while(ptr1!=NULL&&ptr2!=NULL)/*Add
Polynomialsbymerging*/
{if(ptr1>exp>ptr2>exp)
{
if(pol3==NULL)
{
pol3=(structterm*)malloc(size);
pol3>coef=ptr1>coef;
pol3>exp=ptr1>exp;
ptr3=pol3;
ptr1=ptr1>next;
ptr3>next=NULL;
}
else
{
bring=(structterm*)malloc(size);
bring>coef=ptr1>coef;
bring>exp=ptr1>exp;
ptr3>next=bring;
ptr3=bring;
ptr1=ptr1>next;
ptr3>next=NULL;
}
}
else
{
if(ptr1>exp<ptr2>exp)
{
if(pol3==NULL)

{
pol3=(structterm*)malloc(size);
pol3>coef=ptr2>coef;
pol3>exp=ptr2>exp;
ptr3=pol3;
ptr2=ptr2>next;
ptr3>next=NULL;
}
else
{
bring=(structterm*)malloc(size);
bring>coef=ptr2>coef;
bring>exp=ptr2>exp;
ptr3>next=bring;
ptr3=bring;
ptr2=ptr2>next;
ptr3>next=NULL;
}
}
else
{
if(pol3==NULL)
{
pol3=(structterm*)malloc(size);
pol3>coef=ptr1>coef+ptr2>coef;
pol3>exp=ptr2>exp;
ptr3=pol3;
ptr2=ptr2>next;
ptr1=ptr1>next;
ptr3>next=NULL;
}
else
{
bring=(structterm*)malloc(size);
bring>coef=ptr1>coef+ptr2>coef;
bring>exp=ptr2>exp;
ptr3>next=bring;
ptr3=bring;
ptr2=ptr2>next;
ptr1=ptr1>next;
ptr3>next=NULL;

}
}
}
}/*while*/
/*AppendrestofthePolynomial*/
if(ptr2>next==NULL)ptr3>next=ptr1;
if(ptr1>next==NULL)ptr3>next=ptr2;
/*PrinttheFinalPolynomial*/
ptr=pol3;
printf(\nFinalPolynomial);
while(ptr!=NULL)
{
printf(%d^%d,ptr>coef,ptr>exp);
if(ptr>exp!=0)printf(+);
ptr=ptr>next;
}
}
/*Readthetermsofapolynomoialandcreate
thelinkedlist*/
structterm*create_pol()
{
intsize;
structterm*first,*back,*bring;
size=sizeof(structterm);
first=(structterm*)malloc(size);
printf(\ninputPolynomialtermbyterm.Lastterm
withexp=0);
printf(\nEnteratermcoefexp);
scanf(%d%d,&first>coef,&first>exp);
first>next=NULL;
back=first;
while(back>exp!=0)
{
bring=(structterm*)malloc(size);
printf(\nEnteratermcoefexp);
scanf(%d%d,&bring>coef,&bring>exp);
back>next=bring;
back=bring;
}
back>next=NULL;
returnfirst;

}
EXERCISES
1. What is a linear linkedlist?
2. What is the need for a linkedlist? On what account, linked
lists are better than arrays?
3. Write a program that takes a linkedlist pointed by List and traverses
it in such a manner that after the travel the links of visited nodes become
reversed.
4. Write a program to create a linked list and remove all the duplicate
elements of the list.
5. What is a circular linked list? What are its advantages over linear linked
list? Write a program/ algorithm to insert a node at desired position in a
circular linked list.
6. Write a program that takes two ordered linkedlists as input and
merges them into single ordered linked list.
7. What is a doubly linked list? Write program/algorithm for showing the
following operations on a doubly linked list:

Create
Insert
Delete
2. What are the advantages of doubly linked list over singly linkedlist?
3. Write the program that inserts a node in a linkedlist after a
given node.
4. Write the program that inserts a node in a linkedlist before a
given node.
5. Write a program that deletes a given node from a linkedlist.
6. Write a function that deletes an element from a doubly linkedlist.
7. Implement a stack (LIFO) using singly linkedlist.
8. Write a program to implement a queue (FIFO) using singly linked list.
9. How do you detect a loop in a singly linked list? Write a C program for the
same.
10. How do you find the middle of a linked list without counting the nodes?
Write a C program for the same.
11. Write a C program that takes two lists List1 and List2 and compare
them. The program should return the following:
1:ifList1issmallerthanList2
0:ifList1isequaltoList2
1:ifList1isgreaterthanList2

12. Write a function that returns the nth node from the end of a linked
list.
13. Write a program to create a new linear linked list by selecting alternate
element of a given linearlinkedlist.
14. Write an algorithm to search an element from a given linear linked
list.
15. Explain the merits and demerits of static and dynamic memory allocation
techniques.
16. Define a header linkedlist and explain its utility.

7
Trees
CHAPTER OUTLINE

7.1 Introduction

7.2 Basic Terminology

7.3 Binary Trees
7.4 Representation of a Binary Tree
7.5 Types of Binary Trees
7.6 Weighted Binary Trees and Huffman Algorithm
7.7 Dynamic Dictionary Coding
7.1 INTRODUCTION

There are situations where the nature of data is specific and cannot be
represented by linear or sequential data structures. For example, as
shown in Figure 7.1 hierarchical data such as types of computers is nonlinear by nature.

Fig. 7.1 Hierarchical data

Situations such as roles performed by an entity, options available for
performing a task, parentchild relationships demand non-linear
representation. A tourist can go from Mumbai to Goa by any of the
options shown in Figure 7.2.

Fig. 7.2 Options available for travelling from Mumbai to Goa

Similarly, the information about folders and files is kept by an operating
system as a set of hierarchical nodes as shown in Figure 7.3.

Fig. 7.3 The representation of files and folders

Even the phrase structure of a sentence of a language is represented in a
similar fashion (see Figure 7.4).

Fig. 7.4 The phrase structure of a sentence

A critical look at the data representations, shown in figures 7.1
7.4, indicates that the structures are simply upside down trees where the
entity at the highest level represents the root and the other entities at
lower level are themselves roots for subtrees. For example in Figure 7.2,
the entity Travel is theroot of the tree and the options By Sea and By
Road, By Air and By Rail are called child nodes of Travel, the root
node. It may be noted that the entities By Road, By Sea are themselves
roots of subtrees. The nodes which do not have child nodes are called
leaf nodes. For example, the nodes such as By Jetty, By Ship, By Air,
etc. are leaf nodes.
A tree can be precisely defined as a set of 1 or more nodes arranged in a
hierarchical fashion. The node at the highest level is called a root node.
The other nodes are grouped into T1; T2Tk subsets. Each subset itself
is a tree. A tree with zero nodes is called an empty tree.
It may be noted that the definition of a tree is a recursive definition, i.e.,
the tree has been defined in its own terms. The concept can be better
understood by the illustration given in Figure 7.5.
From Figure 7.5, the following components of a tree can be identified:

Fig. 7.5 The tree and its related terms

Root

:A

Child nodes of A

:B, C, D

Sub-Trees of A

:T1, T2, T3

Root of T1

:B

Root of T2

:C

Root of T3

:D

Leaf nodes

:E, G, H, I, J, K, L

Before proceeding to discuss more on trees, let us have a look at basic

terminology used in context of trees as given in Section 7.2.
7.2 BASIC TERMINOLOGY

Parent: An immediate predecessor of a node is called its parent.

Example: D is parent of I and J.
Child: The successor of a node is called its child.
Example: I and J are child nodes of node D.
Sibling: Children of the same parent are called siblings.
Example: B, C, and D are siblings. E, F, and G are also siblings.
Leaf: The node which has no children is called a leaf or terminal node.
Example: I, J, K, L, etc. are leaf nodes.

Internal node: Nodes other than root and leaf nodes are
called internal nodes or non-terminalnodes.
Example: B, C, D and F are internal nodes.
Root: The node which has no parent.
Example: A is root of the tree given in Figure 7.5. Edge: A line from a
parent to its child is called an edge. Example: A-B, A-C, F-K are edges.
Path: It is a list of unique successive nodes connected by edges.
Example: A-B-F-L and B-F-K are paths.
Depth: The depth of a node is the length of its path from root node. The
depth of root node is taken as zero.
Example: The depths of nodes G and L are 2 and 3, respectively.
Height: The number of nodes present in the longest path of the tree
from root to a leaf node is called the height of the tree. This will contain
maximum number of nodes. In fact, it can also be computed as one more
than the maximum level of the tree.
Example: One of the longest paths in the tree shown in Figure 7.3 is A-BF-K. Therefore, the height of the tree is 4.
Degree: Degree of a node is defined as the number of children present
in a node. Degree of a tree is defined equal to the degree of a node with
maximum children.
Example: Degree of nodes C and D are 1 and 2, respectively. Degree of
tree is 3 as there are two nodes A and B having maximum degree equal to
3.
It may be noted that the node of a tree can have any number of child
nodes, i.e., branches. In the same tree, any other node may not have any
child at all. Now the problem is how to represent such a general tree
where the number of branches varies from zero to any possible number.
Or is it possible to put a restriction on number of children and still
manage this important non-linear data structure? The answer is Yes.
We can restrict the number of child nodes to less than or equal to 2. Such
a tree is calledbinary tree. In fact, a general tree can be represented as
a binary tree and this solves most of the problems. A detailed discussion
on binary trees is given in next section.

7.3 BINARY TREES

Binary tree is a special tree. The maximum degree of this tree is 2.

Precisely, in a binary tree, each node can have no more than two
children. A nonempty binary tree consists of the following:
o A node called the root node
o A left sub-tree
o A right sub-tree

Both the sub-trees are themselves binary trees as shown in Figure 7.6.

Fig. 7.6 A binary tree

Kindly note that contrary to a general tree, an empty binary tree can be
empty, i.e., may have no node at all. Examples of valid binary trees are
given in Figure 7.7.
Since the binary trees shown in Figure 7.7 are not connected by a root, it
is called a collection of disjoint trees or a forest. The binary tree shown
in Figure 7.7 (d) is peculiar tree in which there is no right child. Such a
tree is called skewed to the left tree. Similarly, there can be a
tree skewed to the right, i.e., having no left child.

Fig. 7.7 Examples of valid binary trees

There can be following two special cases of binary tree:
1. Full binary tree
2. Complete binary tree

A full binary tree contains the maximum allowed number of nodes at

all levels. This means that each node has exactly zero or two children.

The leaf nodes at the last level will have zero children and the other nonleaf nodes will have exactly two children as shown in Figure 7.8.
A complete binary tree is a full tree except at the last level where all
nodes must appear as far left as possible. A complete binary tree is
shown in Figure 7.9. The nodes have been numbered and there is no
missing number till the last node, i.e., node number 12.

Fig. 7.8 A full binary tree

Fig. 7.9 Complete binary tree

Consider the binary tree given in Figure 7.10 which violates the condition
of the complete binary tree as the node number 11 is missing. Therefore,
it is not a complete binary tree because there is a missing intermediate
node before the last node, i.e., node number 12.

Fig. 7.10 An incomplete binary tree

7.3.1 Properties of Binary Trees
The binary tree is a special tree with degree 2. Therefore, it has some
specific properties which must be considered by the students before
implementing binary trees; these are:
Consider the binary tree given in Figure 7.10.

Fig. 7.10 A full binary tree

1. The maximum number of nodes at level 0 = 1
The maximum number of nodes at level 1 = 2
The maximum number of nodes at level 2 = 4
The maximum number of nodes at level 3 = 8
By induction, it can be deduced that maximum nodes in a binary tree
at level I = 2I
2. The maximum number of nodes possible in a binary tree of depth 0 = 1
The maximum number of nodes possible in a binary tree of depth 1 = 3
The maximum number of nodes possible in a binary tree of depth 2 = 7
The maximum number of nodes possible in a binary tree of depth 3 = 15
By induction, it can be deduced that maximum nodes in a binary tree
at depth k = 2k 1
3. It may be noted that the number of leaf nodes is 1 more than the
number of non-leaf nodes in a non-empty binary tree. For example, a
binary tree of depth 2 has 4 leaf nodes and 3 non-leaf nodes. Similarly, for a
binary tree of depth 3, number of leaf nodes is 8 and non-leaf nodes is 7.
4. Similarly, it can be observed that the number of nodes is 1 more than
the number of edges in a non-empty binary tree, i.e., n = e + 1,
where n is number of nodes and e is number of edges.
5. A skewed tree contains minimum number of nodes in a binary tree.
Consider the skewed tree given in Figure 7.7 (d). The depth of this tree is 3
and the number of nodes in this tree is equal to 4, i.e., 1 more than the
depth. Therefore, it can be observed that the number of nodes in a skewed
binary tree of a depth k is equal to k + 1. In other words, it can be said
that minimum number of nodes possible in a binary tree of depth
k is equal to k +1
7.4 REPRESENTATION OF A BINARY TREE

The binary tree can be represented in two wayslinear representation

and linked representation. A detailed discussion on both the
representations is given in subsequent sections.

7.4.1 linear Representation of a Binary Tree

A binary tree can be represented in a one-dimensional array. Since an
array is a linear data structure, the representation is accordingly called
linear representation of a binary tree. In fact, a binary tree of height h
can be stored in an array of size 2h 1. The representation is quite simple
and follows the following rules:
1.
2.
3.
4.

Store a node at ith location of the array.

Store its left child at (2*i)th location.
Store its right child at (2*i + 1)th location.
The root of the node is always stored at first location of the array.

Consider the tree given in Figure 7.8. Its linear representation using an
array called linTree is given inFigure 7.11.
Fig. 7.11 Linear representation using linTree (of the tree of Figure 7.8)
In an array of size N, for the given node i, the following relationships are
valid in linear representation:
1. leftChild(i)=2*i{When2*i>Nthenthereisnoleft
child}
2. rightChild(i)=2*i+1{When2*i+1>Nthenthereis
norightchild}
3. parent(i)=[i/2]{node0hasnoparentitisaroot
node}

The following examples can be verified from the tree shown in Figure
7.8:
o The parent of node at i = 5 (i.e., E) would be at [5/2] = 2, i.e., B.
o The left child of node at i = 4 (i.e., D) would be at (2*4) = 8, i.e., H.
o The right child of node i = 10 (i.e., J) would be at (2*10) + 1 5 21 which is >
15, indicating that it has no right child.

It may be noted that this representation is memory efficient only when

the binary tree is full whereas in case of skewed trees, this representation
is very poor in terms of wastage of memory locations. For instance,
consider the skewed tree given in Figure 7.7 (d). Its linear representation
using an array called skewTree is given in Figure 7.12.
Fig. 7.12 Linear representation of the skewed tree of Figure 7.7(d) using
skewTree

It may be noted that out of 15 locations, only 4 locations are being used
and 11 are unused resulting in almost 300 per cent wastage of storage
space.
The disadvantages of linear representation of binary tree are as follows:
1. In a skewed or scanty tree, a lot of memory space is wasted.
2. When binary tree is full then space utilization is complete but insertion and
deletion of nodes becomes a cumbersome process.
3. In case of a full binary tree, insertion of a new node is not possible because
an array is a static allocation and its size cannot be increased at run time.

7.4.2 linked Representation of a Binary Tree

A critical look at the organization of a non-empty binary tree suggests
that it is a collection of nodes. Each node comprises the following:
o It contains some information.
o It has an edge to a left child node.
o It has an edge to a right child node.

The above components of a node can be comfortably represented by a

self-referential structure consisting of DATA part, a pointer to its left
child (say leftChild), and a pointer to its right child
(sayrightChild) as shown in Figure 7.13.

Fig. 7.13 A node of a binary tree

Now the node shown in Figure 7.13 will become the building block for
construction of a binary tree as shown in Figure 7.14.

Fig. 7.14 A binary tree represented through linked structures

It may be noted that the root node is pointed by a pointer called binTree.
Both leftChild andrightChild of leaf nodes are pointing to NULL.
Let us now try to create binary tree.
Note: Most of the books create only binary search trees which are very
specific trees and easy to build. In this book a novel algorithm is

provided that creates a generalized binary tree with the help of a stack.
The generalized binary tree would be constructed and implemented in C
so that the reader gets the benefit of learning the technique for building
binary trees.
7.4.2.1 Creation of a Generalized Binary Tree
Let us try to create the tree shown in Figure 7.15.

Fig. 7.15 The binary tree to be created

A close look at the tree shown in Figure 7.15 gives a clue that any binary
tree can be represented as a list of nodes. The list can be prepared by
following steps:
1. Write the root.
2. Write its sub-trees within a set of parenthesis, separated by a comma.
3. Repeat steps 1 and 2 for sub-trees till you reach leaf nodes. A missing child
in a non-leaf node is represented as $.

By applying the above steps on binary tree of Figure 7.15, we get the
following list:

So the final list is: A ( B (D (H, I), E (J, &)), C (F (K,L),G))

Now the above list can be used to create its originating binary tree with
the help of a stack. It is stored in a list called treeList as given below:
treeList[]
=

{A,(,B,(,D,(,H,
I,),E,(,J,$,)
,),C,(,F,(,K,L,
),G,),),#};

The algorithm that uses the treeList and a stack to create the required
binary tree is given below:

Algorithm createBinTree( )
{
Step
1.TakeanelementfromtreeList;
2.If(element!=(&&element!=)&&element!=
$&&element!=
#)
{
Takeanewnodeinptr;
DATA(ptr)=element;
leftChild(ptr)=NULL;
rightChild(ptr)=NULL;
Push(ptr);
}
3.if(element=))
{
Pop(LChild);
Pop(RChild);
Pop(Parent);
leftChild(Parent)=LChild;
rightChild(Parent)=RChild;
Push(Parent);
}
4.if(element=()donothing;
5.if(element=$)Push(NULL);
6.repeatsteps1to5tillelement!=#;
7.Pop(binTree);
8.stop.
}
A simulation of the above algorithm for the following treeList is given
in Figure 7.16.
treeList[]
=

{A,(,B,(,D,(,H,
I,),E,(,J,$,)
,),C,(,F,(,K,
L,),G,),),#};

Fig. 7.16 The creation of binary tree using the algorithm

createBinTree( )
It may be noted that the algorithm developed by us has created exactly
the same tree which is shown inFigure 7.15.
7.4.3 Traversal of Binary Trees
A binary tree is traversed for many purposes such as for searching a
particular node, for processing some or all nodes of the tree and the like.
A system programmer may implement a symbol table as binary search
tree and process it in a search/insert fashion.
The following operations are possible on a node of a binary tree:
1. Process the visited node V.
2. Visit its left child L.
3. Visit its right child R.

The travel of the tree can be done in any combination of the above given
operations. If we restrict ourselves with the condition that left child node
would be visited before right child node, then the following three
combinations are possible:

1. L-V-R : Inorder travel

travel the left sub-tree

process the visited node
travel the right sub-tree
2. V-L-R : Preorder travel

process the visited node

travel the left sub-tree
travel the right sub-tree
2. L-R-V : Postorder travel

travel the left sub-tree

travel the right sub-tree
process the visited node
It may be noted that the travel strategies given above have been named
as preorder, inoder, and postorder according to the placement of
operation: V process the visited node. Inorder (L-V-R) travel has the
operation V in between the other two operations. In preorder (V-L-R),
the operation V precedes the other two operations. Similarly, the
postorder traversal has the operation V after the other two operations.
7.4.3.1 Inorder Travel (L-V-R)
This order of travel, i.e., L-V-R requires that while travelling the binary
tree, the left sub-tree of a node be travelled first before the data of the
visited node is processed. Thereafter, its right sub-tree is travelled.
Consider the binary tree given in Figure 7.17. Its root node is pointed by
a pointer called Tree.

Fig. 7.17 Partial travel of L-V-R order

The travel starts from root node A and its data is not processed as it has a
left child which needs to be travelled first indicated by the operation L.

Therefore, we move to B. As B has no left child, its data is processed

indicated by L the operation V. Now the right child of B needs to be
travelled which is NULL, indicated by the B V operation R. Thus, the
travel of B is complete L and the travel of left sub-tree of A is also over. V
R As per the strategy, the data of A is processed (V) and the travel moves
to its right sub-tree, i.e., to node C (R). The process is repeated for this
sub-tree with root node as C. The final trace of the travel is given below:
BAFDGCE
An algorithm for inorder travel (L-V-R) is given below. It is provided
with the pointer called Tree that points to the root of the binary tree.
Algorithm inorderTravel(Tree)
{
if(Tree==NULL)return
else
{
inorderTravel(leftChild(Tree));
processDATA(Tree);
inorderTravel(rightChild(Tree));
}
}
Example 1: Write a program that creates a generalized binary tree and
travels it using an inorder strategy. While travelling, it should print the
data of the visited node.
Solution: We would use the algorithms createBinTree( )
and inorderTravel(Tree) to write the program.
The required program is given below:

/*Thisprogramcreatesabinarytreeandtravelsit
usinginorderfashion*/
#include<stdio.h>
#include<alloc.h>
#include<conio.h>
structtnode
{
charch;

structtnode*leftChild,*rightChild;
};
structnode
{
structtnode*item;
structnode*next;
};
structnode*push(structnode*Top,structnode
*ptr);
structnode*pop(structnode*Top,structnode
**ptr);
voiddispStack(structnode*Top);
voidtreeTravel(structtnode*binTree);
voidmain()
{inti;
/*chartreeList[]={1,(,2,(,$,
4,
),3,(,5,6,),),#};*/
chartreeList[]={A,(,B,(,D,(,H,
I,),E,(,
J,$,),),C,(,F,(,K,L,),
G,),),#};
structnode*Top,*ptr,*LChild,*RChild,*Parent;
structtnode*binTree;
intsize,tsize;
size=sizeof(structnode);
Top=NULL;
i=0;
while(treeList[i]!=#)
{
switch(treeList[i])
{
case):Top=pop(Top,&RChild);
Top=pop(Top,&LChild);
Top=pop(Top,&Parent);
Parent>item>leftChild=LChild>item;
Parent>item>rightChild=RChild>item;
Top=push(Top,Parent);
break;

case(:break;
default:
ptr=(structnode*)malloc(size);
if(treeList[i]==$)
ptr>item=NULL;
else
{
ptr>item>ch=treeList[i];
ptr>item>leftChild=NULL;
ptr>item>rightChild=NULL;
ptr>next=NULL;
}
Top=push(Top,ptr);
break;
}
i++;
}
Top=pop(Top,&ptr);
binTree=ptr>item;
printf(\nTheTreeis..);
treeTravel(binTree);
}
structnode*push(structnode*Top,structnode
*ptr)
{
ptr>next=Top;
Top=ptr;
returnTop;
}
structnode*pop(structnode*Top,structnode
**ptr)
{
if(Top==NULL)returnNULL;
*ptr=Top;
Top=Top>next;
returnTop;
}
voiddispStack(structnode*Top)
{
if(Top==NULL)
printf(\nStackEmpty);

else
{
printf(\nTheStackis:);
while(Top!=NULL)
{
printf(%c,Top>item>ch);
Top=Top>next;
}
}
}
voidtreeTravel(structtnode*binTree)
{
if(binTree==NULL)return;
treeTravel(binTree>leftChild);
printf(%c,binTree>ch);
treeTravel(binTree>rightChild);
}
It may be noted that a linked stack has been used for the
implementation. The program has been tested for the binary tree given
in Figure 7.15 represented using the following tree list:
treeList[]
=

{A,(,B,(,D,(,H,
I,),E,(,J,$,)
,),C,(,F,(,K,L,),
G,),),#};

The output of the program is given below:

HDIBJEAKFLCG
The above output has been tested to be true.
The program was also tested on the following binary tree (shown as a
comment in this program):
char
treeList[]=

{1,(,2,(,$,4,
),3,(,5,6,),),#}

The output of the program is given below:

241536
It is left as an exercise for the students to generate the binary tree from
the tree list and then verify the output.
7.4.3.2 Preorder Travel (V-L-R)
This order of travel, i.e., V-L-R requires that while travelling the binary
tree, the data of the visited node be processed first before the left subtree of a node is travelled. Thereafter, its right sub-tree is travelled.
Consider the binary tree given in Figure 7.18. Its root node is pointed by
a pointer called Tree.

Fig. 7.18 Partial travel of V-L-R order

The travel starts from root node A and its data is processed (V). As it has
a left child, we move to B (L) and its data is processed (V). As B has no
left child, the right child of B needs to be travelled which is also NULL.
Thus, the travel of B is complete and the processing of L node A and
travel of left sub-tree of A is also over. As per the strategy, the travel
moves to its right sub-tree, i.e., to node C (R). The process is repeated for
this sub-tree with root node as C. The final trace of the travel is given
below:
ABCDFGE
An algorithm for preorder travel (V-L-R) is given below. It is provided
with a pointer called Tree that points to the root of the binary tree.
Algorithm preorderTravel(Tree)
{
if(Tree==NULL)return
else
{processDATA(Tree);
preorderTravel(leftChild(Tree));

preorderTravel(rightChild(Tree));
}
}
Example 2: Write a program that creates a generalized binary tree and
travels it using preorder strategy. While travelling, it should print the
data of the visited node.
Solution: We would use the
algorithms createBinTree() and preorderTravel(Tree) to
write the program. However, the code will be similar to Example1.
Therefore, only the functionpreOrderTravel() is provided.
The required program is given below:
void preOrderTravel(structtnode*binTree)
{
if(binTree==NULL)return;
printf(%c,binTree>ch);
preOrderTravel(binTree>leftChild);
preOrderTravel(binTree>rightChild);
}
The function has been tested for the binary tree given in Figure
7.15 represented using the following tree list:
treeList[]
=

{A,(,B,(,D,(,H,
I,),E,(,J,$,
),),C,(,F,
(,K,L,),G,),),#};

The output of the program is given below:

ABDHIEJCFKLG
The above output has been tested to be true.
The program was also tested on the following binary tree:
char
treeList[]=

{1,(,2,(,$,4,
),3,(,5,6,),),#}

The output of the program is given below:

124356
It is left as an exercise for the students to generate the binary tree from
the treeList and then verify the output.
7.4.3.3 Postorder Travel (L-R-V)
This order of travel, i.e., L-R-V requires that while travelling the binary
tree, the data of the visited node be processed only after both the left and
right sub-trees of a node have been travelled.
Consider the binary tree given in Figure 7.19. Its root node is pointed by
a pointer called Tree.

Fig. 7.19 Partial travel of L-R-V order

The travel starts from root node A and its data is not processed as it has a
left child which needs to be travelled first indicated by the operation L.
Therefore, we move to B. As B has no left child, the right child of B needs
to be travelled which is NULL indicated by the operation R. Now, data of
node B is processed indicated by the operation V. Thus, the travel of B is
complete and the travel of left sub-tree of A is also over. As per the
strategy, the travel moves to its right sub-tree, i.e., to node C (R). The
process is repeated for this sub-tree with root node as C. The final trace
of the travel is given below:
BFGDECA
An algorithm for postorder travel (L-R-V) is given below. It is provided
with the pointer called Tree that points to the root of the binary tree.
Algorithm postOrderTravel(Tree)
{
if(Tree==NULL)return
else
{postOrderTravel(leftChild(Tree));
postOrderTravel(rightChild(Tree));

processDATA(Tree);
}
}
Example 3: Write a program that creates a generalized binary tree and
travels it using preorder strategy. While travelling, it should print the
data of the visited node.
Solution: We would use the
algorithms createBinTree() and postOrderTravel(Tree) to
write the program. However, the code will be similar to Example 1.
Therefore, only the functionpostOrderTravel() is provided.
The required function is given below:
voidpostorderTravel(structtnode*binTree)
{
if(binTree==NULL)return;
postOrderTravel(binTree>leftChild);
postOrderTravel(binTree>rightChild);
printf(%c,binTree>ch);
}
The function has been tested for the binary tree given in Figure
7.15 represented using the following tree list:
treeList[]
=

{A,(,B,(,D,(,H,
I,),E,(,J,$,
),),C,(,F,(,K,L,
),G,),),#};

The output of the program is given below:

HIDJEBKLFGCA
The above output has been tested to be true.
The program was also tested on the following binary tree:
char
treeList[]=

{1,(,2,(,$,4,
),3,(,5,6,),),#}

The output of the program is given below:

425631
It is left as an exercise for the students to generate the binary tree from
the treeList and then verify the output.
7.5 TYPES OF BINARY TREES

There are many types of binary trees. Some of the popular binary trees
are discussed in subsequent sections.
7.5.1 Expression Tree
An expression tree is basically a binary tree. It is used to express an
arithmetic expression. The reason is simple because both binary tree and
arithmetic expression are binary in nature. In binary tree, each node has
two children, i.e., left and right child nodes whereas in an arithmetic
expression, each arithmetic operator has two operands. In a binary tree,
a single node is also considered as a binary tree. Similarly, a variable or
constant is also the smallest expression.
Consider the preorder arithmetic expression given below:
*+A*BC-DE
The above arithmetic expression can be expressed as an expression tree
given in Figure 7.20.

Fig. 7.20 An expression tree of (* + A * B C D E)

The utility of an expression tree is that the various tree travel algorithms
can be used to convert one form of arithmetic expression to another i.e.
prefix to infix or infix to post fix or vice-versa.
For example, the inorder travel of the tree given in Figure 7.20 shall
produce the following infix expression.
A+B*C*DE
The postorder travel of the tree given in Figure 7.20 shall produce the
following postfix expression:

ABC*+DE*
Similarly, the preorder travel of the tree given in Figure 7.20 shall
produce the following prefix expression:
*+A*BCDE
From the above discussion, we can appreciate the rationale behind the
naming of the travel strategies VLR, LVR, and LRV as preorder, inoder,
and postorder, respectively.
Example 4: Write a program that travels the expression tree of Figure
7.21 in an order chosen by the user from a menu. The data of visited
nodes are printed. Design an appropriate menu for this program.

Fig. 7.21 An expression tree

Solution: The expression tree is being expressed as a treeList given
below:
treeList=(+(P,/(Q,A)),B)#
The above list would be used to write the required menu driven program.
The various binary tree travel functions, developed above, would be used
to travel the expression tree as per the user choice entered on the
following menu displayed to the user:

Enter your choice:

The required program is given below:
/*Thisprogramtravelsanexpressiontree*/
#include<stdio.h>

#include<alloc.h>
#include<conio.h>
structtnode
{
charch;
structtnode*leftChild,*rightChild;
};
structnode
{
structtnode*item;
structnode*next;
};
structnode*push(structnode*Top,structnode
*ptr);
structnode*pop(structnode*Top,structnode
**ptr);
voiddispStack(structnode*Top);
voidpreOrderTravel(structtnode*binTree);
voidinOrderTravel(structtnode*binTree);
voidpostOrderTravel(structtnode*binTree);
voidmain()
{inti,choice;
chartreeList[]={,(,+,(,P,/,(,
Q,A,),)
,B,),#};
structnode*Top,*ptr,*LChild,*RChild,*Parent;
structtnode*binTree;
intsize,tsize;
size=sizeof(structnode);
Top=NULL;
i=0;
while(treeList[i]!=#)
{
switch(treeList[i])
{
case):Top=pop(Top,&RChild);
Top=pop(Top,&LChild);
Top=pop(Top,&Parent);
Parent>item>leftChild=LChild>item;
Parent>item>rightChild=RChild>item;

Top=push(Top,Parent);
break;
case(:break;
default:
ptr=(structnode*)malloc(size);
if(treeList[i]==$)
ptr>item=NULL;
else
{
ptr>item>ch=treeList[i];
ptr>item>leftChild=NULL;
ptr>item>rightChild=NULL;
ptr>next=NULL;
}
Top=push(Top,ptr);
break;
}
i++;
}
Top=pop(Top,&ptr);
/*displayMenu*/
do{
clrscr();
printf(\nMenuTreeTravel);
printf(\n\nInorder1);
printf(\nPreorder2);
printf(\nPostorder3);
printf(\nQuit4);
printf(\n\nEnteryourchoice:);
scanf(%d,&choice);
binTree=ptr>item;
switch(choice)
{
case1:printf(\nTheTreeis..);
inOrderTravel(binTree);
break;
case2:printf(\nTheTreeis..);
preOrderTravel(binTree);
break;
case3:printf(\nTheTreeis..);
postOrderTravel(binTree);

break;
};
printf(\nEnteranykeytocontinue);
getch();
}
while(choice!=4);
}
structnode*push(structnode*Top,structnode
*ptr)
{
ptr>next=Top;
Top=ptr;
returnTop;
}
structnode*pop(structnode*Top,structnode
**ptr)
{
if(Top==NULL)returnNULL;
*ptr=Top;
Top=Top>next;
returnTop;
}
voiddispStack(structnode*Top)
{
if(Top==NULL)
printf(\nStackEmpty);
else
{
printf(\nTheStackis:);
while(Top!=NULL)
{
printf(%c,Top>item>ch);
Top=Top>next;
}
}
}
voidpreOrderTravel(structtnode*binTree)
{
if(binTree==NULL)return;
printf(%c,binTree>ch);

preOrderTravel(binTree>leftChild);
preOrderTravel(binTree>rightChild);
}
voidinOrderTravel(structtnode*binTree)
{
if(binTree==NULL)return;
inOrderTravel(binTree>leftChild);
printf(%c,binTree>ch);
inOrderTravel(binTree>rightChild);
}
voidpostOrderTravel(structtnode*binTree)
{
if(binTree==NULL)return;
postOrderTravel(binTree>leftChild);
postOrderTravel(binTree>rightChild);
printf(%c,binTree>ch);
}
The output of the program for choice 1 (inorder travel) is given below:
The Tree is : P + Q / A B
The output of the program for choice 2 (preorder travel) is given below:
The Tree is : + P / Q A B
The output of the program for choice 3 (postorder travel) is given below:
The Tree is : P Q A / + B
It is left as an exercise for the readers to verify the results by manually
converting the expression to the required form.
7.5.2 Binary Search Tree
As the name suggests, a binary search tree (BST) is a special binary tree
that is used for efficient searching of data. In fact a BST stores data in
such a way that binary search algorithm can be applied.
The BST is organized as per the following extra conditions on a binary
tree:
o The data to be stored must have unique key values.

o The key values of the nodes of left sub-tree are always less than the key
value of the root node.
o The key values of the nodes of right sub-tree are always more than the key
value of the root node.
o The left and right sub-trees are themselves BSTs.

In simple words we can say that the key contained in the left child is less
than the key of its parent node and the key contained in the right child is
more than the key of its parent node as shown in Figure 7.22.

Fig. 7.22 A binary search tree (BST)

Kindly note that the inorder travel of the binary search tree would
produce the following node sequence:
Tree is 83 85 90 104 108 109 110 117 125 128
The above list is sorted in ascending order indicating that a binary
search tree is another way to represent and produce a sorted
list. Therefore, a binary search algorithm can be easily applied to
a BST and, hence, the name, i.e., binary search tree.
7.5.2.1 Creation of a Binary Search Tree
A binary search tree is created by successively reading values for the
nodes. The first node becomes the root node. The next nodes value is
compared with the value of the root node. If it is less, the node is
attached to roots left sub-tree otherwise to right sub-tree, and so on.
When the sub-tree is NULL, then the node is attached there itself. An
algorithm for creation of binary search tree is given below.
Note: This algorithm gets a pointer called binTree. The creation
of BST stops when an invalid data is inputted. It uses algorithm attach( )
to attach a node at an appropriate place in BST.
binTree=NULL;/*initiallythetreeisempty*/
AlgorithmcreateBST(binTree)
{
while(1)

{
takeanodeinptr
readDATA(ptr);
if(DATA(ptr)isnotvalid)break;
leftChild(ptr)=NULL;
rightChild(ptr)=NULL;
attach(binTree,ptr);
}
}
returnbinTree;
Algorithmattach(Tree,node)
{
if(Tree==NULL)
Tree=node;
else
{
if(DATA(node)<DATA(Tree)
attach(Tree>leftChild);
else
attach(Tree>rightChild);
}
returnTree;
}
Example 5: Write a program that creates a BST wherein each node of
the tree contains non-negative integer. The process of creation stops as
soon as a negative integer is inputted by the user. Travel the tree using
inorder travel to verify the created BST.
Solution: The algorithm createBST() and its associated
algorithm attach() would be used to write the program. The required
program is given below:

/*ThisprogramcreatesaBSTandverifiesitby
travellingitusinginordertravel*/
#include<stdio.h>
#include<alloc.h>
#include<conio.h>
structbinNode

{
intval;
structbinNode*leftChild,*rightChild;
};
structbinNode*createBinTree();
structbinNode*attach(structbinNode*tree,struct
binNode*node);
voidinOrderTravel(structbinNode*binTree);
voidmain()
{
structbinNode*binTree;
binTree=createBinTree();
printf(\nThetreeis);
inOrderTravel(binTree);
}
structbinNode*createBinTree()
{intval,size;
structbinNode*treePtr,*newNode;
treePtr=NULL;
printf(\nEnterthevaluesofthenodesterminated
byanegativevalue);
val=0;
size=sizeof(structbinNode);
while(val>=0)
{
printf(\nEnterval);
scanf(%d,&val);
if(val>=0)
{
newNode=(structbinNode*)malloc(size);
newNode>val=val;
newNode>leftChild=NULL;
newNode>rightChild=NULL;
treePtr=attach(treePtr,newNode);
}
}
returntreePtr;
}
structbinNode*attach(structbinNode*tree,struct
binNode*node)
{

if(tree==NULL)
tree=node;
else
{
if(node>val<tree>val)
tree>leftChild=attach(tree>leftChild,node);
else
tree>rightChild=attach(tree>rightChild,node);
}
returntree;
}
voidinOrderTravel(structbinNode*binTree)
{
if(binTree==NULL)return;
inOrderTravel(binTree>leftChild);
printf(%d,binTree>val);
inOrderTravel(binTree>rightChild);
}
The above program was given the following input:
56 76 43 11 90 65 22
And the following output was obtained when this BST was travelled
using inorder travel:
11 22 43 56 65 76 90
It may be noted that the output is a sorted list verifying that the
created BST is correct. Similarly, the program was given the following
input:
45 12 34 67 78 56
and the following output was obtained:
12 34 45 56 67 78
The output is a sorted list and it verifies that the created BST is correct.

Fig. 7.23 The skewed binary search tree

Note: The drawback of a binary search tree is that if the data supplied to
the creation process are in a sequence or order then the BST would be a
skewed tree. The reason being that every succeeding number being
greater than its preceding number would become a right child and would
result in a skewed right binary tree. For example, the following input
would produce a skewed tree shown inFigure 7.23.
25 35 43 67 89 120
7.5.2.2 Searching in a Binary Search Tree
The most desirable and important application of a BST is that it is used
for searching an entry in a set of values. This method is faster as
compared to arrays and linked lists. In an array, the list of items needs to
be sorted before a binary search algorithm could be applied for searching
an item into the list. On the contrary, the data automatically becomes
sorted during the creation of a BST. The item to be searched is compared
with the root. If it matches with the data of the root, then success is
returned otherwise left or right sub-tree is searched depending upon the
item being less or more than the root.
An algorithm for searching a BST is given below; it searches Val in
a BST pointed by a pointer calledTree.
Algorithm searchBST(Tree,Val)
{
if(Tree=NULL)
returnfailure;
if(DATA(Tree)==val)
returnsuccess;
else
if(Val<DATA(Tree))
search(Tree>leftChild,Val);
else
search(Tree>rightChild,Val);

}
Example 6: Write a program that searches a given key in a binary
search tree through a function called searchBST(). The function
returns 1 or 0 depending upon the search being successful or a failure.
The program prompts appropriate messages like Search successful or
Search unsuccessful.
Solution: The above algorithm searchBST() is used. The required
program is given below:
#include<stdio.h>
#include<alloc.h>
#include<conio.h>
structbinNode
{
intval;
structbinNode*leftChild,*rightChild;
};
structbinNode*createBinTree();
structbinNode*attach(structbinNode*tree,struct
binNode*node);
intsearchBST(structbinNode*binTree,intval);
voidmain()
{intresult,val;
structbinNode*binTree;
binTree=createBinTree();
printf(\nEnterthevaluetobesearched);
scanf(%d,&val);
result=searchBST(binTree,val);
if(result==1)
printf(\nSearchSuccessful);
else
printf(\nSearchUnSuccessful);
}
structbinNode*createBinTree()
{}

/*thisfunctioncreateBinTree()issameasdeveloped
inaboveexamples*/
intsearchBST(structbinNode*binTree,intval)
{intflag;
if(binTree==NULL)
return0;
if(val==binTree>val)
return1;
else
if(val<binTree>val)
flag=searchBST(binTree>leftChild,val);
else
flag=searchBST(binTree>rightChild,val);
returnflag;
}
The above program has been tested.
7.5.2.3 Insertion into a Binary Search Tree
The insertion in a BST is basically a search operation. The item to be
inserted is searched within theBST. If it is found, then the insertion
operation fails otherwise when a NULL pointer is encountered the item is
attached there itself. Consider the BST given in Figure 7.24.
It may be noted that the number 49 is inserted into the BST shown
in Figure 7.24. It is searched in theBST which comes to a dead end at the
node containing 48. Since the right child of 48 is NULL, the number 49
is attached there itself.
Note: In fact, a system programmer maintains a symbol table as
a BST and the symbols are inserted into the BST in search-insert fashion.
The main advantage of this approach is that duplicate entries into the
table are automatically caught. The search engine also stores the
information about downloaded web pages in search-insert fashion so
that possible duplicate documents from mirrored sites could be caught.
An algorithm for insertion of an item into a BST is given below:
Algorithm insertBST(binTree,item)
{
if(binTree==NULL)
{
Takenodeinptr;

DATA(ptr)=item;
leftChild(ptr)=NULL;
rightChild(ptr)=NULL;
binTree=ptr;
returnsuccess;
}
else
if(DATA(binTree)==item)
returnfailure;
else
{Takenodeinptr;
DATA(ptr)=item;
leftChild(ptr)=NULL;
rightChild(ptr)=NULL;
if(DATA(binTree)>item)
insertBST(binTree>leftChild,ptr);
else
insertBST(binTree>rightChild,ptr);
}
}

Fig. 7.24 Insertion operation in BST

Example 7: Write a program that inserts a given value in a binary
search tree through a function called insertBST(). The function

returns 1 or 0 depending upon the insertion being success or a failure.

The program prompts appropriate message like Insertion successful or
Duplicate value. Verify the insertion by travelling the tree in inorder
and displaying its contents containing the inserted node.
Solution: The above algorithm insertBST() is used. The required
program is given below:
/*ThisprograminsertsavalueinaBST*/
#include<stdio.h>
#include<alloc.h>
#include<conio.h>
structbinNode
{
intval;
structbinNode*leftChild,*rightChild;
};
structbinNode*createBinTree();
structbinNode*attach(structbinNode*tree,struct
binNode*node);
structbinNode*insertBST(structbinNode*binTree,
intval,int*flag);
voidinOrderTravel(structbinNode*Tree);
voidmain()
{intresult,val;
structbinNode*binTree;
binTree=createBinTree();
printf(\nenterthevaluetobeinserted);
scanf(%d,&val);
binTree=insertBST(binTree,val,&result);
if(result==1)
{printf(\nInsertionSuccessful,TheTreeis..);
inOrderTravel(binTree);
}
else
printf(\nduplicatevalue);
}
structbinNode*createBinTree()

/*thisfunctioncreateBinTree()issameasdeveloped
inaboveexamples*/
{}
structbinNode*attach(structbinNode*tree,struct
binNode*node)
{
if(tree==NULL)
tree=node;
else
{
if(node>val<tree>val)
tree>leftChild=attach(tree>leftChild,node);
else
tree>rightChild=attach(tree>rightChild,node);
}
returntree;
}
structbinNode*insertBST(structbinNode*binTree,
intval,int
*flag)
{intsize;
structbinNode*ptr;
size=sizeof(structbinNode);
if(binTree==NULL)
{ptr=(structbinNode*)malloc(size);
ptr>val=val;
ptr>leftChild=NULL;
ptr>rightChild=NULL;
binTree=ptr;
*flag=1;
returnbinTree;
}
if(val==binTree>val)
{
*flag=0;
returnbinTree;
}
else
if(val<binTree>val)
binTree>leftChild=insertBST(binTree>leftChild,val,
flag);

else
binTree>rightChild=insertBST(binTree>rightChild,
val,flag);
returnbinTree;
}
voidinOrderTravel(structbinNode*Tree)
{
if(Tree!=NULL)
{
inOrderTravel(Tree>leftChild);
printf(%d,Tree>val);
inOrderTravel(Tree>rightChild);
}
}
The above program was tested for the following input:
Nodes of the BST: 45 12 67 20 11 56
Value to be inserted: 11
The output: Duplicate value was obtained indicating that the value is
already present in the BST. Similarly, for the following input:
Nodes of BST: 65 78 12 34 89 77 22
Value to be inserted: 44
The output: Insertion successful, the Tree is .. 12 22 34 44 65 77 78 89,
indicating that the insertion at proper place in BST has taken place.
7.5.2.4 Deletion of a Node from a Binary Search Tree
Deletion of a node from a BST is not as simple as insertion of a node. The
reason is that the node to be deleted can be at any of the following
positions in the BST:
1. The node is a leaf node.
2. The node has only one child.
3. The node is an internal node, i.e., having both the children.

Case 1: This case can be very easily handled by setting the pointer to the
node from its parent equal toNULL and freeing the node as shown
in Figure 7.25.

Fig. 7.25 Deletion of a leaf node

The leaf node containing 49 has been deleted. Kindly note that its
parents pointer has been set to NULLand the node itself has been freed.
Case 2: This case can also be handled very easily by setting the pointer
to the node from its parent equal to its child node. The node may be
freed later on as shown in Figure 7.26.
The node, containing 48, having only one child (i.e., 49) has been
deleted. Kindly note that its parents pointer (i.e., 50) has been set to its
child (i.e., 49) and the node itself has been freed.
Case 3: From discussions on BST, it is evident that when a BST is
travelled by inorder, the displayed contents of the visited nodes are
always in increasing order, i.e., the successor is always greater then the
predecessor. Another point to be noted is that inorder successor of an
internal node (having both children) will never have a left child and it is
always present in the right sub-tree of the internal node.

Fig. 7.26 Deletion of a node having only one child

Consider the BST given in Figure 7.26 The inorder successor of internal
node 45 is 48, and that of 56 is 58. Similarly, in Figure 7.22, the inorder
successors of internal nodes 90, 108, and 110 are 104, 109, and 117,
respectively and each one of the successor node has no left child. It may
be further noted that the successor node is always present in the right
sub-tree of the internal node.

The successor of a node can be reached by first moving to its right child
and then keep going to left till aNULL is found as shown in Figure 7.27.
It may be further noted that if an internal node ptr is to be deleted then
the contents of inorder successor of ptr should replace the contents
of ptr and the successor be deleted by the methods suggested for Case 1
or Case 2, discussed above . The mechanism is shown in Figure 7.28.
An algorithm to find the inorder successor of a node is given below. It
uses a pointer called succPtrthat travels from ptr to its successor and
returns it.
Algorithm findSucc(ptr)
{
succPtr=rightChild(ptr);
while(leftChild(succPtr!=NULL))
succPtr=leftChild(succptr);
returnsuccPtr;
}

Fig. 7.27 Finding inorder successor of a node

Fig. 7.28 Deletion of an internal node in BST

The algorithm that deletes a node from a BST is given below. It takes
care of all the three cases given above.
Algorithm delNodeBST(Tree,Val)
{
ptr=search(Tree,Val);/*Findsthenodetobe
deleted*/
parent=search(Tree,Val);/*Findstheparentof
thenode*/
if(ptr==NULL)reporterrorandreturn;
if(leftChild(ptr)==NULL&&rightChild(ptr)==
NULL)/*Case1*/
{
if(leftChild(Parent)==ptr)
leftChild(Parent)=NULL;
else
if(rightChild(Parent)==ptr)
rightChild(Parent)=NULL;
free(ptr);
}

if(leftChild(ptr)!=NULL&&rightChild(ptr)==
NULL)/*Case2*/
{
if(leftChild(Parent)==ptr)
leftChild(Parent)=leftChild(ptr);
else
if(rightChild(Parent)==ptr)
rightChild(Parent)=leftChild(ptr);
free(ptr);
}
else
if(leftChild(ptr)==NULL&&rightChild(ptr)!=
NULL)/*Case2*/
{
if(leftChild(Parent)==ptr)
leftChild(Parent)=rightChild(ptr);
else
if(rightChild(Parent)==ptr)
rightChild(Parent)=rightChild(ptr);
free(ptr);
}
if(leftChild(ptr)!=NULL&&rightChild(ptr)!=
NULL)/*Case3*/
{
succPtr=findSucc(ptr);
DATA(ptr)=DATA(succPtr);
delNodeBST(Tree,succPtr);
}
Example 8: Write a program that deletes a given node form binary
search tree. The program prompts appropriate message like Deletion
successful. Verify the deletion for all the three cases by travelling the
tree in inorder and displaying its contents.
Solution: The above algorithms
del NodeBST() and findSucc() are used. The required program is
given below:
/*ThisprogramdeletesanodefromaBST*/
#include<stdio.h>
#include<alloc.h>

#include<conio.h>
structbinNode
{
intval;
structbinNode*leftChild,*rightChild;
};
structbinNode*createBinTree();
structbinNode*attach(structbinNode*tree,struct
binNode*node);
structbinNode*searchBST(structbinNode*binTree,
intval,int*flag);
structbinNode*findParent(structbinNode*binTree,
structbinNode*ptr);
structbinNode*delNodeBST(structbinNode*binTree,
intval,int*flag);
structbinNode*findSucc(structbinNode*ptr);
voidinOrderTravel(structbinNode*Tree);
voidmain()
{intresult,val;
structbinNode*binTree;
binTree=createBinTree();
printf(\nEnterthevaluetobedeleted);
scanf(%d,&val);
binTree=delNodeBST(binTree,val,&result);
if(result==1)
{printf(\nDeletionsuccessful,TheTreeis..);
inOrderTravel(binTree);
}
else
printf(\nNodenotpresentinTree);
}
structbinNode*createBinTree()/*thisfunction
createBinTree()issameasdevelopedinabove
examples*/
{
}
structbinNode*attach(structbinNode*tree,struct
binNode*node)
{/*thisfunctionattach()issameasdevelopedin
aboveexamples*/

}
structbinNode*delNodeBST(structbinNode*binTree,
intval,int*flag)
{intsize,nval;
structbinNode*ptr,*parent,*succPtr;
if(binTree==NULL)
{*flag=0;
returnbinTree;
}
ptr=searchBST(binTree,val,flag);
if(*flag==1)
parent=findParent(binTree,ptr);
else
returnbinTree;
if(ptr>leftChild==NULL&&ptr>rightChild==
NULL)/*Case1*/
{
if(parent>leftChild==ptr)
parent>leftChild=NULL;
else
if(parent>rightChild==ptr)
parent>rightChild=NULL;
free(ptr);
}
if(ptr>leftChild!=NULL&&ptr>rightChild==
NULL)
{
if(parent>leftChild==ptr)
parent>leftChild=ptr>leftChild;
else
if(parent>rightChild==ptr)
parent>rightChild=ptr>leftChild;
free(ptr);
returnbinTree;
}
else
if(ptr>leftChild==NULL&&ptr>rightChild!=
NULL)
{
if(parent>leftChild==ptr)
parent>leftChild=ptr>rightChild;

else
if(parent>rightChild==ptr)
parent>rightChild=ptr>rightChild;
free(ptr);
returnbinTree;
}
if(ptr>leftChild!=NULL&&ptr>rightChild!=
NULL)
{
succPtr=findSucc(ptr);
nval=succPtr>val;
delNodeBST(binTree,succPtr>val,flag);
ptr>val=nval;
}
returnbinTree;
}
voidinOrderTravel(structbinNode*Tree)
{
if(Tree!=NULL)
{
inOrderTravel(Tree>leftChild);
printf(%d,Tree>val);
inOrderTravel(Tree>rightChild);
}
}
structbinNode*findSucc(structbinNode*ptr)
{structbinNode*succPtr;
getch();
succPtr=ptr>rightChild;
while(succPtr>leftChild!=NULL)
succPtr=succPtr>leftChild;
returnsuccPtr;
}
structbinNode*searchBST(structbinNode*binTree,
intval,int*flag)
{
if(binTree==NULL)
{*flag=0;
returnbinTree;
}

else
{
if(binTree>val==val)
{
*flag=1;
returnbinTree;
}
else
{if(val<binTree>val)
returnsearchBST(binTree>leftChild,val,flag);
else
returnsearchBST(binTree>rightChild,val,flag);
}
}
}
structbinNode*findParent(structbinNode*binTree,
structbinNode*ptr)
{structbinNode*pt;
if(binTree==NULL)
returnbinTree;
else
{
if(binTree>leftChild==ptr||binTree>rightChild
==ptr)
{pt=binTree;returnpt;}
else
{if(ptr>val<binTree>val)
pt=findParent(binTree>leftChild,ptr);
else
pt=findParent(binTree>rightChild,ptr);
}
}
returnpt;
}
The above program was tested for the following input:
Case 1:
Nodes of the BST: 38 27 40 11 39 30 45 28
Value to be deleted: 11, a leaf node

The output: Deletion successful-The Tree is.. 27 28 30 38 39 40

45 was obtained.
Case 2:
Similarly, for the following input:
Nodes of BST: 38 27 40 11 39 30 45 28
Value to be deleted: 30, a node having no right child.
The output: Deletion successful, the Tree is.. 11 27 28 38 39 40 45,
indicating that the non-leaf node 30 having only one child has been
deleted.
Case 3:
Similarly, for the following input:
Nodes of BST: 38 27 40 11 39 30 45 28
Value to be deleted: 40, an internal node having both the children.
The output: Deletion successful, the Tree is.. 11 27 28 30 38 39 45,
indicating that the internal node 40 having no child has been deleted.
7.5.3 Heap Trees
A heap tree is a complete binary tree. This data structure is widely used
for building priority queues. Sorting is another application in which heap
trees are employed; the sorting method is known as heap sort. As the
name suggests, a heap is a collection of partially ordered objects
represented as a binary tree. The data at each node in heap tree is less
than or equal to the data contained in its left child and the right child as
shown in Figure 7.29. This property of each node having data part less
than or equal to its children is called heap-order property.

Fig. 7.29A min-heap tree

As we go from leaf nodes to the root (see Figure 7.29), data of nodes
starts decreasing and, therefore, the root contains the minimum data
value of the tree. This type of heap tree is called a min heap.

On the contrary, if a node contains data greater than or equal to the data
contained in its left and right child, then the heap is called as max
heap The reason being that as we move from leaf nodes to the root, the
data values increase and, therefore, in a max heap the data in the root is
the maximum data value in the tree as shown in Figure 7.30.

Fig. 7.30 A max-heap tree

A complete binary tree is called a heap tree (say max heap) if it has the
following properties:
1. Each node in the heap has data value more than or equal to its left and right
child nodes.
2. The left and right child nodes themselves are heaps.

Thus, every successor of a node has data value less than or equal to the
data value of the node. The height of a heap with N nodes = |log 2 N|.
7.5.3.1 Representation of a Heap Tree
Since a heap tree is a complete binary tree, it can be comfortably and
efficiently stored in an array. For example, the heap of Figure 7.30 can be
represented in an array called heap as shown in Figure 7.31. The zeroth
position contains total number of elements present in the heap. For
example, the zeroth location of heap contains 10 indicating that there are
10 elements in the heap shown in Figure 7.31.
Fig. 7.31 Linear representation of a heap tree
It may be noted that this representation is simpler than the linked
representation as there are no links. Moreover from any node of the tree,
we can move to its parent or to its children, i.e., in both forward and
backward directions. As discussed in Section 7.4.1, the following
relationships hold good:
In an array of size N, for the given node i, the following relationships are
valid in linear representation:
(1) leftChild (i) = 2*i

{When 2*i > N then there is no left child}

(2) rightChild(i) = 2*i +

1

{When 2*i +1 > N then there is no right

child}

(3) parent (i) = | i/2 |

{node 0 has no parent it is a root node}

The operations that are defined on a heap tree are:

1. Insertion of a node into a heap tree
2. Deletion of a node from a heap tree

A discussion on these operations is given in the subsequent sections.

7.5.3.2 Insertion of a Node into a Heap Tree
The insertion operation is an important operation on a heap tree. The
node to be inserted is placed on the last available location in the heap,
i.e., in the array. Its data is compared with its parent and if it is more
than its parent, then the data of both the parent and the child are
exchanged. Now the parent becomes a child and it is compared with its
own parent and so on; the process stops only when the data part of child
is found less than its parent or we reach to the root of the heap. Let us try
to insert 69 in the heap given in Figure 7.31. The operation is illustrated
in Figure 7.32.
An algorithm that inserts an element called item into a heap represented
in an array of size N calledmaxHeap is given below. The number of
elements currently present in the heap is stored in the zeroth location
of maxHeap.

Fig. 7.32 Insertion of a node into a heap tree

Algorithm insertHeap(item)
{
if(heap[0]==0)
{heap[0]=1;
heap[1]=item;
}
else
{
last=heap[0]+1;
if(last>size)
promptmessageHeapFull;
else
{
heap[last]=item;
heap[0]=heap[0]+1;
I=last;
Flag=1;
while(Flag)
{
J=abs(I/2);/*findparentsindex*/
if(J>=1)
{

if(heap[J]<heap[I])
{
temp=heap[J];
heap[J]=heap[I];
heap[I]=temp;/*Parentbecomeschild*/
I=J;
}
else
Flag=0;
}
else
Flag=0;
}
}
}
}
}
Note: The insertion operation is important because the heap itself is
created by iteratively inserting the various elements starting from an
empty heap.
Example 9: Write a program that constructs a heap by iteratively
inserting elements input by a user into an empty heap.
Solution: The algorithm insertHeap() would be used. The required
program is given below:
/*Thisprogramcreatesaheapbyiteratively
insertingelementsintoaheapTree*/
#include<stdio.h>
#include<conio.h>
voidinsertHeap(intheap[],intitem,intsize);
voiddispHeap(intheap[]);
voidmain()
{
intitem,i,size;
intmaxHeap[20];
printf(\nEnterthesize(<20)ofheap);
scanf(%d,&size);
printf(\nEntertheelementsofheaponebyone);
for(i=1;i<=size;i++)

{
printf(\nElement:);
scanf(%d,&item);
insertHeap(maxHeap,item,size);
}
printf(\nTheheapis..);
dispHeap(maxHeap);
}
voidinsertHeap(intheap[],intitem,intsize)
{intlast,I,J,temp,Flag;
if(heap[0]==0)
{heap[0]=1;
heap[1]=item;
}
else
{
last=heap[0]+1;
if(last>size)
{
printf(\nHeapFull);
}
else
{
heap[last]=item;
heap[0]++;
I=last;
Flag=1;
while(Flag)
{
J=(int)(I/2);
if(J>=1)
{/*findparent*/
if(heap[J]<heap[I])
{
temp=heap[J];
heap[J]=heap[I];
heap[I]=temp;
I=J;
}
else
Flag=0;

}
else
Flag=0;
}
}
}
}
voiddispHeap(intheap[])
{
inti;
for(i=1;i<=heap[0];i++)
printf(%d,heap[i]);
}
The above program was tested for the input: 22 15 56 11 7 90 44
The output obtained is: 90 15 56 11 7 22 44, which is found to be
correct.
In the next run, 89 was added to the list, i.e., 22 15 56 11 7 90 44 89.
We find that the item 89 has been inserted at proper place as indicated
by the output obtained thereof: 90 89 56 15 7 22 44 11
It is left as an exercise for the reader to verify above results.
7.5.3.3 Deletion of a Node from a Heap Tree
The deletion operation on heap tree is relevant when its root is deleted.
Deletion of any other element has insignificant applications. The root is
deleted by the following steps:
Step
1.
2.
3.
4.

Save the root in temp.

Bring the right most leaf of the heap into root.
Move root down the heap till the heap is ordered.
Return temp containing the deleted node.

The operation, given in step2 above is also called reheap operation.

Consider the heap given in Figure 7.33.

Fig. 7.33 A heap tree

Let us delete its root (i.e., 70) by bringing the rightmost leaf (i.e., 56) into
the root as shown in Figure 7.34. The root has been saved in temp.
An algorithm that deletes the maximum element (i.e., root) from a heap
represented in an array of size N called maxHeap is given below. The
number of elements currently present in the heap is stored in the zeroth
location of maxHeap.

Fig. 7.34 Deletion of root from a heap

Algorithm delMaxHeap(heap)
{
if(heap[0]==1)returnheap[1];/*Thereisonly
oneelementintheheap*/
last=heap[0];
tempVal=heap[1];/*Savetheitemtobedeletedand
returned*/
heap[1]=heap[last];/*Bringthelastleafnodeto
root*/
heap[0];/*Reducethenumberofelementsby1*/
K=last1;
I=1;/*Pointtoroot*/

Flag=1;
while(Flag)
{
J=I*2;
if(J<=K)
{/*findbiggerchildandstoreitspositioninpos
*/
if(heap[J]>heap[J+1])
pos=J;
else
pos=J+1;
if(heap[pos]>heap[I])/*performreheap*/
{
temp=heap[pos];
heap[pos]=heap[I];
heap[I]=temp;
I=pos;/*parentbecomesthechild*/
}
else
Flag=0;
}
else
Flag=0;
}
returntempVal;
}
Example 10: Write a program that deletes and displays the root of a
heap. The remaining elements of the heap are put to reheap operation.
Solution: The algorithm delMaxHeap() would be used. The required
program is given below:
#include<stdio.h>
#include<conio.h>
intdelMaxHeap(intheap[]);
voidinsertHeap(intheap[],intitem,intsize);
voiddispHeap(intheap[]);
voidmain()
{
intitem,i,size;
intmaxHeap[20];

printf(\nEnterthesize(<20)ofheap);
scanf(%d,&size);
printf(\nEntertheelementsofheaponebyone);
for(i=1;i<=size;i++)
{
printf(\nElement:);
scanf(%d,&item);
insertHeap(maxHeap,item,size);
}
printf(\nTheheapis..);
dispHeap(maxHeap);
item=delMaxHeap(maxHeap);
printf(\nThedeleteditemis:%d,item);
printf(\nTheheapafterdeletionis..);
dispHeap(maxHeap);
}
voidinsertHeap(intheap[],intitem,intsize)
{intlast,I,J,temp,Flag;
if(heap[0]==0)
{heap[0]=1;
heap[1]=item;
}
else
{
last=heap[0]+1;
if(last>size)
{
printf(\nHeapFull);
}
else
{
heap[last]=item;
heap[0]++;
I=last;
Flag=1;
while(Flag)
{
J=(int)(I/2);
if(J>=1)
{/*findparent*/
if(heap[J]<heap[I])

{
temp=heap[J];
heap[J]=heap[I];
heap[I]=temp;
I=J;
}
else
Flag=0;
}
else
Flag=0;
}
}
}
}
voiddispHeap(intheap[])
{
inti;
for(i=1;i<=heap[0];i++)
printf(%d,heap[i]);
}
intdelMaxHeap(intheap[])
{
inttemp,I,J,K,val,last,Flag,pos;
if(heap[0]==1)returnheap[1];
last=heap[0];
val=heap[1];
heap[1]=heap[last];
heap[0];
K=last1;
I=1;
Flag=1;
while(Flag)
{
J=I*2;
if(J<=K)
{/*findchild*/
if(heap[J]>heap[J+1])
pos=J;
else
pos=J+1;

if(heap[pos]>heap[I])
{
temp=heap[pos];
heap[pos]=heap[I];
heap[I]=temp;
I=pos;
}
else
Flag=0;
}
else
Flag=0;
}
returnval;
}
The above program was tested for input: 34 11 56 78 23 89 12 5
and the following output was obtained:
The heap is: 89 56 78 11 23 34 12 5
The deleted item is: 89
The heap after deletion is: 78 56 34 11 23 5 12
It is left as an exercise for the students to verify the results by manually
making the heaps before and after deletion of the root.
7.5.3.4 Heap Sort
From discussion on heap trees, it is evident that the root of
a maxHeap contains the largest value. When the root of the max heap is
deleted, we get the largest element of the heap. The remaining elements
are put to reheap operation and we get a new heap. If the two operations
delete and reheap operation is repeated till the heap becomes empty and
the deleted elements stored in the order of their removal, then we get a
sorted list of elements in descending order as shown in Figure 7.35.

Fig. 7.35 Repetitive deletion from a heap produces sorted list

Example 11: Write a program that sorts a given list of numbers by heap
sort.
Solution: We would modify the program of Example10 wherein the
element deleted from heap would be stored in a separate array
called sortList. The deletion would be carried out iteratively till the
heap becomes empty. Finally, sortList would be displayed. The
required program is given below:
/*Thisprogramsortsagivenlistofnumbersusinga
heaptree*/
#include<stdio.h>
#include<conio.h>
intdelMaxHeap(intheap[]);
voidinsertHeap(intheap[],intitem,intsize);
voiddispHeap(intheap[]);
voidmain()
{
intitem,i,size,last;
intmaxHeap[20];
intsortList[20];
printf(\nEnterthesize(<20)ofheap);
scanf(%d,&size);
printf(\nEntertheelementsofheaponebyone);
for(i=1;i<=size;i++)

{
printf(\nElement:);
scanf(%d,&item);
insertHeap(maxHeap,item,size);
}
printf(\nTheheapis..);
dispHeap(maxHeap);
getch();
last=maxHeap[0];
for(i=0;i<last;i++)
{
item=delMaxHeap(maxHeap);
sortList[i]=item;
}
printf(\nThesortedlist);
for(i=0;i<last;i++)
printf(%d,sortList[i]);
}
voidinsertHeap(intheap[],intitem,intsize)
{intlast,I,J,temp,Flag;
if(heap[0]==0)
{heap[0]=1;
heap[1]=item;
}
else
{
last=heap[0]+1;
if(last>size)
{
printf(\nHeapFull);
}
else
{
heap[last]=item;
heap[0]++;
I=last;
Flag=1;
while(Flag)
{
J=(int)(I/2);
if(J>=1)

{/*findparent*/
if(heap[J]<heap[I])
{
temp=heap[J];
heap[J]=heap[I];
heap[I]=temp;
I=J;
}
else
Flag=0;
}
else
Flag=0;
}
}
}
}
voiddispHeap(intheap[])
{
inti;
for(i=1;i<=heap[0];i++)
printf(%d,heap[i]);
}
intdelMaxHeap(intheap[])
{
inttemp,I,J,K,val,last,Flag,pos;
if(heap[0]==1)returnheap[1];
last=heap[0];
val=heap[1];
heap[1]=heap[last];
heap[0];
K=last1;
I=1;
Flag=1;
while(Flag)
{
J=I*2;
if(J<=K)
{/*findchild*/
if(heap[J]>heap[J+1])

pos=J;
else
pos=J+1;
if(heap[pos]>heap[I])
{
temp=heap[pos];
heap[pos]=heap[I];
heap[I]=temp;
I=pos;
}
else
Flag=0;
}
else
Flag=0;
}
returnval;
}
The above program was tested for the input: 87 45 32 11 34 78 10 31
The output obtained is given below:
The heap is: 87 45 78 31 34 32 10 11
The sorted list is 87 78 45 34 32 31 11 10
Thus, the program has performed the required sorting operation using
heap sort.
Note: The above program is effective and gives correct output but it is
inefficient from the storage point of view. The reason being that it is
using an extra array called sortList to store the sorted list.
The above drawback can be removed by designing an in-place heap
sort algorithm that uses the same array to store the sorted list in which
the heap also resides. In fact, the simple modification could be that let
root and the last leaf node be exchanged. Without disturbing the last
node, reheap the remaining heap. Repeat the exchange and reheap
operation iteratively till only one element is left in the heap. Now, the
heap contains the elements in ascending order. The mechanism is
illustrated inFigure 7.36.

Fig. 7.36 The in-place heap sort

Algorithm IpHeapSort()
{
if(heap[0]>1)
{
do
{
last=heap[0];
temp=heap[1];/*exchangerootandlastleaf
node*/
heap[1]=heap[last];
heap[last]=temp;
heap[0];
last=heap[0];/*lastpointstonewlastleaf
node*/
K=last1;
I=1;
Flag=1;
while(Flag)
{
J=I*2;
if(J<=K)
{/*findchildwhichisgreaterthanparent*/
if(heap[J]>heap[J+1])
pos=J;
else

pos=J+1;
if(heap[pos]>heap[I])
{
temp=heap[pos];/*exchangeparentandchild
*/
heap[pos]=heap[I];
heap[I]=temp;
I=pos;
}
else
Flag=0;
}
else
Flag=0;
}
}
while(last>1);
}
}
It may be noted that the element at the root is being exchanged with the
last leaf node and the rest of the heap is being put to reheap operation.
This is iteratively being done in the do-while loop till there is only one
element left in the heap.
Example 12: Write a program that in-place sorts a given list of
numbers by heap sort.
Solution: We would employ the algorithm IpHeapSort() wherein the
element from the root is exchanged with the last leaf node and rest of the
heap is put to reheap operation. The exchange and reheap is carried out
iteratively till the heap contains only 1 element.
The required program is given below:

/*Thisprograminplacesortsagivenlistof
numbersusingheaptree*/
#include<stdio.h>
#include<conio.h>
voidsortMaxHeap(intheap[]);
voidinsertHeap(intheap[],intitem,intsize);
voiddispHeap(intheap[]);

voidmain()
{
intitem,i,size,last;
intmaxHeap[20];
printf(\nEnterthesize(<20)ofheap);
scanf(%d,&size);
printf(\nEntertheelementsofheaponebyone);
for(i=1;i<=size;i++)
{
printf(\nElement:);
scanf(%d,&item);
insertHeap(maxHeap,item,size);
}
printf(\nTheheapis..);
dispHeap(maxHeap);
getch();
sortMaxHeap(maxHeap);
printf(\nThesortedlist);
for(i=1;i<=size;i++)
printf(%d,maxHeap[i]);
}
voidinsertHeap(intheap[],intitem,intsize)
{intlast,I,J,temp,Flag;
if(heap[0]==0)
{heap[0]=1;
heap[1]=item;
}
else
{
last=heap[0]+1;
if(last>size)
{
printf(\nHeapFull);
}
else
{
heap[last]=item;
heap[0]++;
I=last;
Flag=1;
while(Flag)

{
J=(int)(I/2);
if(J>=1)
{/*findparent*/
if(heap[J]<heap[I])
{
temp=heap[J];
heap[J]=heap[I];
heap[I]=temp;
I=J;
}
else
Flag=0;
}
else
Flag=0;
}
}
}
}
voiddispHeap(intheap[])
{
inti;
for(i=1;i<=heap[0];i++)
printf(%d,heap[i]);
}
voidsortMaxHeap(intheap[])
{
inttemp,I,J,K,val,last,Flag,pos;
if(heap[0]>1)
{
do
{
last=heap[0];
temp=heap[1];/*Exchangerootwiththelastnode
*/
heap[1]=heap[last];
heap[last]=temp;
heap[0];
last=heap[0];

K=last1;
I=1;
Flag=1;
while(Flag)
{
J=I*2;
if(J<=K)
{/*findchild*/
if(heap[J]>heap[J+1])
pos=J;
else
pos=J+1;
if(heap[pos]>heap[I])
{
temp=heap[pos];
heap[pos]=heap[I];
heap[I]=temp;
I=pos;
}
else
Flag=0;
}
else
Flag=0;
}
}
while(last>1);
}
}
7.5.3.5 Merging of Two Heaps
Merging of heap trees is an interesting application wherein two heap
trees are merged to produce a tree which is also a heap. The merge
operation is carried out by the following steps:
Given two heap trees: Heap1 and Heap2
Step
1. Delete a node from Heap2.
2. Insert the node into Heap1.
3. Repeat steps 1 and 2 till Heap 2 is not empty.

Consider the merge operation shown in Figure 7.37.

Fig. 7.37 Merging of two heap trees

Example 13: Write a program that merges two heap trees to produce a
third tree which is also a heap.
Solution: The required program is given below:
/*Thisprogrammergestwoheaptrees*/
#include<stdio.h>
#include<conio.h>
intdelMaxHeap(intheap[]);
voidinsertHeap(intheap[],intitem,intsize);
voiddispHeap(intheap[]);
voidmain()
{
intitem,i,size1,size2;
intmaxHeap1[20],maxHeap2[20];
printf(\nEnterthesize(<20)ofheap1);
scanf(%d,&size1);
printf(\nEntertheelementsofheaponebyone);
for(i=1;i<=size1;i++)
{
printf(\nElement:);
scanf(%d,&item);
insertHeap(maxHeap1,item,size1);
}
printf(\nTheheap1is..);
dispHeap(maxHeap1);
printf(\nEnterthesize(<20)ofheap2);
scanf(%d,&size2);

printf(\nEntertheelementsofheaponebyone);
for(i=1;i<=size2;i++)
{
printf(\nElement:);
scanf(%d,&item);
insertHeap(maxHeap2,item,size2);
}
printf(\nTheheap2is..);
dispHeap(maxHeap2);
for(i=1;i<=size2;i++)
{
item=delMaxHeap(maxHeap2);
size1=size1+1;
insertHeap(maxHeap1,item,size1);
}
printf(\nTheheapsaftermergeis..);
dispHeap(maxHeap1);
}
voidinsertHeap(intheap[],intitem,intsize)
{intlast,I,J,temp,Flag;
if(heap[0]==0)
{heap[0]=1;
heap[1]=item;
}
else
{
last=heap[0]+1;
if(last>size)
{
printf(\nHeapFull);
}
else
{
heap[last]=item;
heap[0]++;
I=last;
Flag=1;
while(Flag)
{
J=(int)(I/2);
if(J>=1)

{/*findparent*/
if(heap[J]<heap[I])
{
temp=heap[J];
heap[J]=heap[I];
heap[I]=temp;
I=J;
}
else
Flag=0;
}
else
Flag=0;
}
}
}
}
voiddispHeap(intheap[])
{
inti;
for(i=1;i<=heap[0];i++)
printf(%d,heap[i]);
}
intdelMaxHeap(intheap[])
{
inttemp,I,J,K,val,last,Flag,pos;
if(heap[0]==1)returnheap[1];
last=heap[0];
val=heap[1];
heap[1]=heap[last];
heap[0];
K=last1;
I=1;
Flag=1;
while(Flag)
{
J=I*2;
if(J<=K)
{/*findchild*/
if(heap[J]>heap[J+1])
pos=J;

else
pos=J+1;
if(heap[pos]>heap[I])
{
temp=heap[pos];
heap[pos]=heap[I];
heap[I]=temp;
I=pos;
}
else
Flag=0;
}
else
Flag=0;
}
returnval;
}
The above program was tested on the heap trees given in Figure 7.37.
The input provided is given below:
Heap1: 67 31 34 8 10 5
Heap2: 22 12 17 4 9 615
The output obtained is given below:
The heap after merge is 67 31 34 17 12 6 22 8 15 10 9 5 4
It is left for the reader to verify the output from Figure 7.37.
Note: Both MaxHeap and MinHeap trees can be merged. The type of
output merged tree will depend upon the type
of Heap1. If Heap1 is maxHeap, then the heap obtained after merge
operation would also be of type maxHeap. Similarly, if Heap1 is
minHeap, then the heap obtained after merge operation would also be of
type minHeap.
7.5.4 Threaded Binary Trees
From the discussion on binary trees, it is evident that creation of a tree is
a one time process. But, a tree is travelled numerous times consuming a
large part of the overall time taken by an algorithm. Moreover, the travel
algorithms are recursive in nature amounting to large usage of storage.

Therefore, there is a scope for reducing the travels or the number of

steps required to travel a tree. For example, during inorder travel, we
move from a node to its successor and from the successor to its
successor, and so on.
The questions is can we move directly to an inorder successor of a node
in a tree. The answer is yes if we can keep pointers from the nodes to
their successors and move through the tree using these pointers.
We know that in a binary tree of N nodes, there are N +
1 NULL pointers. Therefore, the unused NULLpointers of the binary tree
can be employed for the purpose of non-recursive travelling within a
binary tree. Let us modify the structure of the node of binary tree as per
the following rules:
o If a node is having an empty left child (i.e., a NULL pointer), then replace it
by a special pointer that points to its predecessor node in the tree.
o If a node is having an empty right child (i.e., a NULL pointer), then replace it
by a special pointer that points to its successor node in the tree.
o The special pointer is called a thread.

Consider the binary tree given in Figure 7.38 wherein a binary tree has
been assigned the threads. It may be noted that the pointers have been
shown with thick lines and the threads with dotted lines.

Fig. 7.38 The vacant NULL pointers have been replaced by threads
The resultant binary tree of Figure 7.38 is called a threadedbinary
tree(TBT). It may be noted that the characteristic of this special tree
is that a left thread of a node points to its inorder predecessor and the
right thread to its inorder successor. For example, the left thread of E is
pointing to B, its predecessor and the right thread of E is pointing to
A, its successor.

A threadedbinarytree is a binary tree in which a node uses its

empty left child pointer to point to its inorder predecessor and the empty
right child pointer to its inorder successor.
Now, the problem is that how a program would differentiate between a
pointer and a thread. A simple solution is to keep variables
called LTag and RTag, having value 1 or 0 indicating a pointer or a
thread, respectively. Accordingly, a node of a threaded binary tree would
have the structure as given in Figure 7.39.

Fig. 7.39 A node of a threaded binary tree

Let us now use the structure of Figure 7.39 to represent the threaded
binary tree of Figure 7.38. The arrangement is shown in Figure 7.40.

Fig. 7.40 The threaded binary tree with node structure having the Tag
bits
A left thread pointing to NULL indicates that there is no more inorder
predecessor. Similarly, a NULLright thread indicates that there is no
more inorder successor. For example, the left thread of H is pointing
to NULL and the right thread of G is pointing to NULL. This anomalous
situation can be handled by keeping a head node having a left child and a
right thread as shown in Figure 7.41. This dummy node does not contain
any value. In fact, it is an empty threaded binary tree because both its
left child and right thread are pointing to itself.

Fig. 7.41 A head node indicating an empty threaded binary tree

Now the dangling threads of a binary tree can be made to point to the
head node and the left child of the head node to the root of the binary
threaded tree as shown in Figure 7.42.

Fig. 7.42 A binary threaded tree with a head node

Now with the arrangement shown in Figure 7.42, it can be appreciated
that an inorder successor of a node pointed by ptr can be very easily
found by the following set of rules:
o If RTag of ptr is 0, then the rightChild is pointing to its successor.
o If RTag of ptr is 1 (i.e., thread), then the left most node of its right sub-tree
is its successor.

An algorithm that returns inorder successor of a node is given below:

Algorithm findSucc(ptr)
{
if(RTag(ptr)==0)
returnrightChild(ptr);
else
{
ptr=rightChild(ptr)
while(LTag(ptr)=1)
ptr=leftChild(ptr);
}
returnptr;
}
Similarly, an inorder predecessor of a node pointed by ptr can be very
easily found by the following set of rules:
o If LTag of ptr is 0, then the leftChild is pointing to its predecessor.
o If LTag of ptr is 1 (i.e., thread), then the right most node of its left sub-tree
is its predecessor.

An algorithm that returns inorder successor of a node is given below:

Algorithm findPred(ptr)
{
if(LTag(ptr)==0)
returnleftChild(ptr);
else
{
ptr=leftChild(ptr)
while(RTag(ptr)=1)
ptr=rightChild(ptr);
}
returnptr;
}
Now we can write an algorithm for non-recursive inorder travel of a
threaded binary tree (TBT), which is given below:
Algorithm inOrderTBT(Tree)
{
Tree=Head;
while(1)
{
tree=findSucc(tree);
if(tree=Head)return;
else
printDATA(Tree);
}
}
The following operations can be defined on a threaded binary tree.
o Insertion of a node into a TBT.
o Deletion of a node from a TBT.

As the threaded binary tree has been designed to honour inorder travel,
the insertion and deletion have to be carried out in such a way that after
the operation the tree remains the inorder threaded binary tree. A brief
discussion on both the operations is given in the subsequent sections.
7.5.4.1 Insertion into a Threaded Binary Tree
A node X, having Data part equal to T, can be inserted into an inorder
threaded binary tree after node Y in either of the following four ways:

Case 1: When X is inserted as left child of Y and Y has an empty left child.
Case 2: When X is inserted as right child of Y and Y has an empty right
child.
Case 3: When X is inserted as left child of Y and Y has a non-empty left
child
Case 4: When X is inserted as right child of Y and Y has a non-empty
right child.
CASE 1:
When Y (Say Data part equal to F) has an empty left child, then it must
be a thread. The insertion can be made by making the left thread of X
(Data part equal to T) to point where the left thread of Y is currently
pointing to. Thereafter, X becomes the left child of Y for which LTag of Y
is reset to 1. The right child of X is set as thread pointing to the node Y,
indicating that Y is the successor of X. The insertion operation is shown
in Figure 7.43.

Fig. 7.43 Insertion of a node X after Y having no left child

It may be noted that the node containing T has become the left child of
node containing F. After the insertion, the tree is still a threaded binary
tree. The predecessor of T is A and the successor of T is F.
An algorithm for insertion of node as left thread is given below:
Algorithm insertAtLThread(X,Y)

{
if(LTag(Y)==0)
{
LTag(X)=0;
leftChild(X)=leftChild(Y);
LTAg(Y)=1;
leftChild(Y)=X;
RTag(X)=0;
rightChild(X)=Y;
}
}
CASE 2:
When Y (Data part equal to F) has an empty right child, then it must be
a thread. The insertion can be made by making the right thread of X
(Data part equal to T) to point where the right thread of Y is currently
pointing to. Thereafter, X becomes the right child of Y for which RTag of
Y is reset to 1. The left child of X is set as a thread pointing to the node Y,
indicating that Y is the predecessor of X. The insertion operation is given
in Figure 7.44.

Fig. 7.44 Insertion of a node X after Y having no right child

It may be noted that the node containing T has become the right child
of node containing F. After the insertion, the tree is still a threaded
binary tree. The predecessor of T is F and the successor of T is C.
An algorithm for insertion of node as right thread is given below:
Algorithm inserAtRThread(X,Y)
{
if(RTag(Y)==0)
{
RTag(X)=0;
rightChild(X)=rightChild(Y);
RTAg(Y)=1;
rightChild(Y)=X;
LTag(X)=0;
leftChild(X)=Y;
}
}
CASE 3:
When Y (Data part equal to C) has non-empty left child, the insertion
can be made by making the left child of X (Data part equal to T) point
where the left child of Y is currently pointing to. Thereafter, X becomes
the left child of Y for which Lchild of Y is pointed to X. The right child
of X is set as thread pointing to the node Y, indicating that Y is the
successor of X.
Now the most important step is to find the inorder predecessor of Y
(pointed by ptr) and make it inorder predecessor of X by making ptrs
right thread to point to X. The insertion operation is shown in Figure
7.45.
It may be noted that the node containing T has become the left child of
node containing C. After the insertion, the tree is still a threaded binary
tree. The predecessor of T is F and the successor of T is C.
An algorithm for insertion of node as left child is given below:
Algorithm insertAtLchild(X,Y)
{
RTag(X)=0;/*ConnecttherightthreadofXtoY
*/
rightChild(X)=Y
LTAG(X)=1;/*ConnecttheleftChildofXtothe
nodepointedby

leftchildoY*/
leftChild(X)=leftChild(Y)
/*PointtheleftchildofYtoX*/
leftChild(Y)=X;
ptr=inOrderPred(Y)/*Findtheinorder
predecessorofY*/
rightChild(ptr)=X;/*Connecttherightthread
of(ptr)predecessortoX*/
}
CASE 4:
When Y (Data part equal to C) has non-empty right child, the insertion
can be made by making the right child of X (Data part equal to T) point
where the right child of Y is currently pointing to. Thereafter, X becomes
the right child of Y by pointing the right child of Y to X. The left thread of
X is set as a thread pointing to the node Y, indicating that Y is the
predecessor of X.

Fig. 7.45 Insertion of a node at non empty left child of a given node
Now, the most important step is to find the inorder successor of Y
(pointed by ptr) and make it inorder successor of X by making ptrs left
thread point to X. The insertion operation is shown inFigure 7.46.

It may be noted that the node containing T has become the right child
of node containing C. After the insertion, the tree is still a threaded
binary tree. The predecessor of T is C and the successor of T is G.
An algorithm for insertion of node as right child is given below:
Algorithm insertAtRchild(X,Y)
{
LTag(X)=0;/*ConnecttheleftthreadofXtoY
*/
leftChild(X)=Y
RTag(X)=1;/*ConnecttheleftChildofXtothe
nodepointedbyleftchildoY*/
rightChild(X)=rightChild(Y)
/*PointtheleftchildofYtoX*/
rightChild(Y)=X;
ptr=inOrderSucc(Y)/*Findtheinorder
successorofY*/
leftChild(ptr)=X;/*Connecttheleftthreadof
ptr(successor)toX*/
}

Fig. 7.46 Insertion of a node at non-empty right child of a given node

A brief discussion on deletion of a node from a threaded binary tree is

given in Section 7.5.4.2.
7.5.4.2 Deletion from a Threaded Binary Tree
A node X, having Data part equal to T, can be deleted from an inorder
threaded binary tree in any of the following five ways:
Case 1: When X is a left leaf node.
Case 2: When X is a right leaf node
Case 3: When X is only having a right sub-tree.
Case 4: When X is only having a left sub-tree.
Case 5: When X is having both sub-trees.
Before an attempt is made to delete a node from any tree, the parent of
the node must be reached first so that the dangling pointers produced
after deletion can be properly handled. Therefore, the following
algorithm is given that finds the parent Y of a node X.
Y=Tree;
Algorithm findParent(X,Y)
{
If(Y=NULL)return;
if((Y>leftChild==X||Y>(rightChild==X))
returnY
findParent(X,Y>leftChild);
findParent(X,Y>rightChild);
}
The above algorithm is recursive. A non-recursive algorithm that uses
the threads to reach to the parent of a node X is given below. It uses the
strategy wherein it reaches the successor of the rightmost node. If the
left child of the successor points to X, then the successor is parent.
Otherwise, move to the left child and then keep going to right till a node
is found whose right child is equal to X.
Algorithm findParent(X)
{
ptr=X;
while(RTag(ptr)==1)/*Findthenodewhichhasa
rightthread*/
ptr=rightChild(ptr);
succ=rightChild(ptr);/*Movetothesuccessor
node*/

if(succ>leftChild==X)
Y=succ;/*Parentfound,pointitbyY*/
else
{suc=leftChild(succ);
while(rightChild(succ)!=X)
succ=rightChild(succ);
Y=succ;
}
returnY;
}
Now the algorithm findParent() can be conveniently used to delete a
node from a threaded binary tree. The various cases of deletion are
discussed below:
CASE 1:
When X is a left leaf node, then its parent Y is found (see Figure 7.47).
The left child of Y is made to point where Xs left thread is currently
pointing to. The LTag of Y is set to 0, indicating that it has become a
thread now.
An algorithm for deletion of a node X which is a left leaf node is given
below:
Algorithm delLeftLeaf(X,Y)
{
Y=findParent(X);
if(leftChild(Y)==X)
{
leftChild(Y)=leftChild(X);/*Makeleft
childofparentasthreadandletitpointwhereXs
leftthreadispointing*/
LTag(Y)=0;
}
returnX;
}
It may be noted from Figure 7.47 that node X has been deleted and the
left child of its parent Y has been suitably handled.

Fig. 7.47 Deletion of a left leaf node

CASE 2:
When X is a right leaf node, then its parent Y is found. The right child of
Y is made to point where Xs right thread is currently pointing to.
The RTag of Y is set to 0, indicating that it has become a thread now.
An algorithm for deletion of a node X which is a right leaf node is given
below:
Algorithm delRightLeaf (X,Y)
{
Y=findParent(X);
if(rightChild(Y)==X)
{
rightChild(Y)=rightChild(X);/*Makeright
childofparentasthreadandletitpointwhereXs
leftthreadispointing*/
RTag(Y)=0;
}
returnX;

}
This case is similar to CASE 1 and, therefore, illustration for this case is
not provided.
CASE 3:
When X is having a right sub-tree, its parent Y is found. A pointer called
grandson is make to point to the right child of X (see fig 7.48). The
right child of Y is pointed to grandson. The successor of X is found. The
left thread of successor is pointed to Y, indicating that the successor of X
has now become the successor of parent as shown in Figure 7.48.

Fig. 7.48 Deletion of a node having only right sub-tree

It may be noted that in Figure 7.48, the grandson is also the successor of
X.
An algorithm for deletion of X when X is having a right sub-tree is given
below:
Algorithm delNonLeafR(X,Y)
{

Y=findParent(X);
if(RTag(X)==1)
{
grandson=rightChild(X);
rightChild(Y)=grandson;
succ=inOrderSucc(X);
leftChild(succ)=Y;
}
returnX
}
CASE 4:
When X is having only left sub-tree, its parent Y is found. A pointer
called grandson is made to point to the left child of X. The left child of Y
is pointed to grandson. The predecessor of X is found. The right thread
of predecessor is pointed to Y, indicating that the predecessor of X has
now become the predecessor of parent.
An algorithm for deletion of X when X is having a left sub-tree is given
below:
Algorithm delNonLeafL(X,Y)
{
Y=findParent(X);
if(LTag(X)==1)
{
grandson=leftChild(X);
leftChild(Y)=grandson;
pred=inOrderPred(X);
rightChild(pred)=Y;
}
returnX
}
This case is similar to CASE3 and, therefore, illustration for this case is
not provided.
CASE 5:
When X is having both sub-trees, then X is not deleted but its successor
is found. The data of successor is copied in X and the successor is deleted
using either CASE1orCASE2, whichever is applicable.

An algorithm for deletion of X when X is having both sub-trees is given

below:
Algorithm delBoth(X)
{
succ=findSucc(X);
DATA(X)=DATA(succ);
DeletesuccusingCase1orCase2
}
Note: The applications of threaded binary trees are limited in the sense
that the above discussions pertain to inorder threaded binary trees only.
For postorder, the above algorithms need to be rewritten. The
implementation is rather complex.
7.6 WEIGHTED BINARY TREES AND HUFFMAN ALGORITHM

Information is an important resource but transportation of right

information to a rightful receiver is equally important. For example, a
specialist doctor in Delhi must know the complete online data of a
medical tourist (say from Afghanistan) from the patients hospital in
Kabul.
A search engine collects information from various Web sites with the
help of downloaders calledcrawlers. The information, if large, needs to
be compressed before sending it over network so as to save bandwidth
traffic. The huge data collected by the search engine needs to be encoded
prior to storing it in its local repository.
Similarly, Fax machines must also compress and encoded data before
throwing it on the telephone line.
All these applications require efficient encoding techniques. In this
section, Huffman coding technique is discussed with a view to introduce
the reader to the area of data compression, a potential area of research
nowadays.
The following terms of binary trees A are important for discussions
related to the topic of this section.
2-Tree or Extended Binary Tree: It is binary tree with the property
that each node has either 0 children or 2 children. The tree given
in Figure 7.49 is an Extended Binary Tree (EBT)
Internal Node: The node with two children is called an internal node.
For example, in Figure 7.49, the nodes B, C, D, E , and F are internal
nodes.

External Node: A node with no children is called an external node.

This is also popularly known as aleaf node. For example, the nodes G, H,
I, J, K, L, and M of Figure 7.49 are external or leaf nodes.

Fig. 7.49 An extended binary tree

Path Length: The number of edges traversed to reach a node is called
the path length of the node. For example the path length of Node D is 2
and that of Node K is 3. Of course, the path length of root node A is 0.
As per convention, it may be noted that internal and external nodes have
been represented by circle and rectangle, respectively. This
representation of a binary tree is also called as Extended Binary Tree
(EBT). Thus, in EBT, an internal node is represented as a circle and the
external node as a rectangle.
External Path Length: Let LEXT be the external path length of a binary
tree. It is defined as sum of path lengths of all external nodes. The
external path length of the tree given in Figure 7.49 is given below:
LEXT = LH + LI + LJ + LK + LL + LM + LG = 3 +3 +3 + 3 + 3 +3 + 2 =20
Where Li : is the path length of external node i
This can be formally defined as:
Where Li : is the path length of external node i
Internal Path length: Let LINT be the internal path length of a binary
tree. It is defined as sum of path lengths of all internal nodes. The
internal path length of the tree given in Figure 7.49 is given below:
LINT = LA + LB + LC + LD + LE + LF = 0 +1 +1 + 2 + 2 + 2 =8
Where Li : is the path length of internal node i
This can be formally defined as:
Where Li : is the path length of internal node i

From induction , we can say that : LEXT = LINT + 2 * N, where N is number

of internal nodes.
In EBT of Figure 7.49, N = 6, therefore, LEXT = LINT + 2 * N = 8 + 2 * 6 =
20.
Weighted Binary Tree: So far we have considered the cost of an edge
between two nodes as one unit. In real world the edges may differ in
terms of length, weight, degree of comfort, profit, utility etc. However, in
order to keep things simpler, let us keep the cost between each edge as
one A but assign a weight Wi to an ith external node as shown in Figure
7.50.

Fig. 7.50 A weighted binary tree

Since the weighted binary tree is also an extended binary tree, therefore
it is also called as a weighted extended binary tree.
Let Wi be the weight of an ith external node with path length Li then the
weighted external path length of weighted binary tree Lw is defined as:
Where N is the number of external nodes.
From above given relation, the weighted path length of the extended
binary tree of Figure 7.50 is computed as:
Lw = 5 * 3 + 7 * 3 + 2 * 3 + 6 *3 + 1 * 3 + 4 * 3 + 3 * 2 = 81
The purpose of above discussion was to create a weighted binary tree
with minimum weighted path length. It has many applications such as:
data compression techniques for sending messages over communication
networks, storing of collection of documents on secondary storage, etc.
D. Huffman provided a solution to this problem. The Huffman algorithm
is discussed in next section.
7.6.1 Huffman Algorithm
For a given list of weights w1, w2, w3, , wn, find a weighted binary tree
T of minimum-weighted path length out of all the possible weighted

binary trees. Huffman gave a bottom up approach for this problem. The
simple steps for solving this problem are given below:
1. Sort the given list of weights into increasing order, place them on a priority
queue called List.
2. Pick two minimum weights from the List as external nodes and form a subtree
3. Compute the weighted path list of the sub-tree obtained in Step 2. Insert
this weight in the List
4. Repeat steps 23 till only one weight is left on the List.

Consider the external weights of the tree given in Figure 7.50. The
external nodes with their weights are:
G H I J K L M
3 5 7 2 6 1 4
The sorted list is given below:
L J G M H K I
1 2 3 4 5 6 7
The trace of above given steps is provided in Figure 7.51:

Fig. 7.51 Construction of a Huffman Tree

A copy of the final Huffman tree obtained in Figure 7.51 is given
in Figure 7.52

Fig. 7.52 The final Huffman Tree

Now the weighted path length of this tree can be computed by adding the
weights of all internal nodes as given below:
Lw = 28 + 16 + 9 + 12 + 6 + 3 = 74
It is left as an exercise to the reader to create another weighted binary
tree from the list of weights and verify that the Huffman tree obtained
above has the minimum weighted path length.
It may be noted that when duplicate values are available in the list of
weights then depending upon the order of their positions in the list, we
may have many different Huffman trees. For instance, in the example
given above in second step if we swap the values of 3* and 3 and in third
step, swap the values of ent Huffman tree as compared to the Huffman
tree obtained above (see Figure 7.53). However, both the Huffman trees
will have the same weighted path lengths, i.e. 74.
Thus, construction of Huffman tree by the Huffman algorithm may not
always result in a unique tree.

Fig. 7.53 An alternate Huffman Tree

As discussed above, the Huffman tree can be usefully applied to data

compression techniques for storing files and sending messages. Infact,
Huffman not only gave the algorithm for construction of Huffman tree
but also offered a coding technique for data compression applications.
The technique is discussed in next section.
7.6.2 HUFFMAN CODES

Normally, the alphabets are stored using standard codes such as ASCII,
UNICODE, EBCDIC etc. These are fixed length codes. The ASCII coding
scheme uses 8 bits for a character whereas the UNICODE uses 16 bits.
These codes are useful only when the text, to be stored, contains all the
alphabets with equal number of their instances in the text, otherwise it
becomes a costly method in terms of storage. For example, the string
madam would use 40 bits of storage for using ASCII code and 80 bits
for UNICODE.
However, a closer look reveals that the string consists of 2 instances of
alphabet m, 2 instances of alphabet a, and 1 instance of alphabet d.
Let us now assign them the following binary codes:
Alphabet

Binary code

00

01

10

Using the above codes, we can compress the string madam into a much
shorter code of 10 bits as shown below:
1001000110
Thus, we have saved a significant amount of storage space i.e. 10 bits as
compared to 40 or 80 bits. But the coding scheme used by us is still a
fixed length coding scheme.
Huffman provided an excellent variable length coding scheme which
effectively compresses the data by assigning short codes to most
frequently occurring characters. The less frequent characters get longer
codes. Thus, the technique saves the storage space in terms of 2590 per
cent of the original size of the text. For a given piece of text, the following
steps are used for developing the codes for various alphabets contained
in the text:
1. Count the frequency of occurrence of each alphabet in the text.
2. Based on the frequency, assign weights to the alphabets.

3. Create a Huffman tree using the weights assigned to the alphabets. Thus,
the alphabets become the external nodes of the Huffman tree.
4. Starting from root, label the left edge as 0 and right as 1.
5. Traverse the sequence of edges from root to an external node i.e. an
alphabet. The assembly of labels of the edges from root to the alphabet
becomes the code for the alphabet.

The Huffman codes applied to Huffman tree of Figure 7.52 are given
in Figure 7.54

Fig. 7.54 The assignment Huffman codes

The codes assigned to various leaf nodes are listed in the table given
below:
Character

Weight

00

11

011

010

101

1001

1000

It may be noted that the heaviest nodes: I and K have got shorter codes
than the lighter nodes.
Example 14: Develop Huffman code for this piece of text Nandini

Solution: The trace of steps followed for assigning the Huffman codes
to the various characters of the given text are given below:
1. The frequency count of each character appearing in the above text is given
in the following table:
Character

Frequency

2. Let the frequency be the weight of each character.

3. The sorted list of the characters in the order of their weights is given below:
Nadni
11122
4. From the weights develop the Huffman Tree. The trace of Huffman
algorithm is given Figure 7. 55.

Fig. 7.55 Development of the Huffman tree

5. Assign the Huffman codes as shown in Figure 7.56.

Fig. 7.56 Assignment of Huffman codes

The codes assigned to various leaf nodes are listed in the table given
below:
Character

Weight

00

010

011

10

11

Now, the code assigned to the text Nandini is: 0100111000111011 =

16 bits
In ASCII coding scheme, we would have used 8*7 = 56 bits. Thus, we
have saved 40 bits (56 16) of storage space in this small piece of text
comprising of only 7 characters.
Example 15: Develop Huffman code for this piece of text given below:
when you are on the left you are on the right. when you are on the right,
you are on the wrong
Also compute the amount storage space in terms of bits saved by using
Huffman coding.
Solution: The text contains 93 characters. We will represent the space
with the character: B. The trace of steps followed for assigning the
Huffman codes to the various characters of the given text are given
below:

Step 1: The frequency count of each character appearing in the above

text is given in the following table:

2. Let the frequency be the weight of each character.

3. The sorted list of the characters in the order of their weights is given
below:
4. From the weights, develop the Huffman tree with the help of Huffman
algorithm. The trace is given in Appendix D.
5. Assign the Huffman codes as shown in Figure 7.57.

Fig. 7.57 Assignment of Huffman codes

The codes assigned to various leaf nodes are listed in the table given
below:

As expected, the most frequently occurring character called space () has

got the shortest code i.e. 00. Similarly, the character e has been
assigned a shorter code i.e. 100. Obviously, the least frequently occurring
characters like l & f have been assigned the longest codes 1101000 and
1101001 respectively.
Let us now compute the total space occupied by the text using Huffman
coding. The computed bits per character are given in table below:

Total bits = 348

It may be noted that in ASCII coding scheme, we would have used 8 93
= 744 bits. Thus, we have saved 396 bits (744 348) of storage space in
this piece of text comprising of only 93 characters.
Therefore percentage of space saved = 396 / 744 * 100 = 53.2%
7.7 DYNAMIC DICTIONARY CODING

Dynamic dictionary coding is a common data compression technique. In

this method, the input text is scanned from the first word in such a
manner that:

o The words are inserted into an empty dictionary in a search/insert fashion,

i.e., the word is searched in the dictionary, and if not found it is inserted in
the dictionary and also written on the output.
o If the word is found in the dictionary then its position from the dictionary is
written on the output instead of the word.

Example: Input Text: When you are on the leftyou are on the right
and when you are on the right you are on the wrong.
The dynamic dictionary obtained from above input text is given below:
The dynamic dictionary
Word

position

When

you

are

on

the

left

right

and

Output text: When you are on the left 2 3 4 5 right and 1 2 3 4 5 7 2 3 4

5 wrong.
Example 16. Write a program that reads text into an array of strings
and compresses the data using dynamic dictionary coding technique.
Solution: An array of strings to store 50 words of 20 characters each
would be used to represent the text. The required program is given
below:
/*ThisprogramimplementsDynamicDictionary
Decoding*/
#include<stdio.h>
main()

{
chartext[50][20];
charDictionary[30][20];
inti,j,flag,wordCount,DictCount;
i=1;
printf(\nEnterthetextterminatedbya###:\n)
;
do
{
i++;
scanf(%s,text[i]);
}
while(strcmp(text[i],###));
wordCount=i;
strcpy(Dictionary[0],text[0]);
DictCount=0;/*TheDictionarygetsthefirst
word*/
printf(\nTheTextis:%s,text[0]);/*printthe
firstword*/
for(i=1;i<=wordCount;i++)
{
flag=0;
for(j=0;j<=DictCount;j++)
{
if(!strcmp(text[i],Dictionary[j]))
{
printf(%d,j+1);
flag=1;
break;
}
}/*EndofjLoop*/
if(flag==0)
{
DictCount++;
strcpy(Dictionary[DictCount],text[i]);
printf(%s,text[i]);
}
}/*EndiLoop*/
}
The sample output of the program is given below:

EXERCISES
1. What is the need of tree data structure? Explain with the help of examples.
List out the areas in which this data structure can be applied extensively.
2. What is a tree? Discuss why definition of tree is recursive. Why it is said to
be non-linear?
3. Define the following terms:

i.
ii.
iii.
iv.

Root
Empty tree
Leaf nodes
Sub-tree

2. Discuss the following terms with suitable examples:

i.
ii.
iii.
iv.
v.
vi.
vii.
viii.
ix.
2.
3.
4.
5.

6.

Parent
Child
Sibling
Internal node
Edge
Path
Depth
Height
Degree

Explain binary tree with the help of examples. Discuss the properties of
binary tree that need to be considered.
Discuss the concept of full binary tree and complete binary tree.
Differentiate between the two types with the help of examples.
Discuss various methods of representation of a binary tree along with their
advantages and disadvantages.
Present an algorithm for creation of a binary tree. Consider a suitable binary
tree to be created and discuss how the algorithm can be applied to create the
required tree.
What do you mean by tree traversal? What kinds of operations are possible
on a node of a binary tree? Give an algorithm for inorder traversal of a
binary tree. Taking an example, discuss how the binary tree can be
traversed using inorder traversal.

7.
8.

9.
10.

What do you mean by preorder traversal of a tree? Discuss an algorithm for

preorder traversal of a binary tree along with an example.
Give an algorithm for postorder traversal of a binary tree. Taking an
example, discuss how the binary tree can be traversed using postorder
traversal.
What are the various types of binary trees?
What is an expression tree? Consider the following preorder arithmetic
expressions:

i.
ii.
iii.
iv.
v.
vi.

2.

3.

4.

5.

6.

Draw expression tree for the above arithmetic expressions. Give the inorder,
preorder and postorder traversal of the expression trees obtained.
What is a binary search tree? What are the conditions applied on a binary
tree to obtain a binary search tree? Present an algorithm to construct a
binary search tree and discuss it using a suitable example.
Discuss various operations that can be applied on a BST. Present an
algorithm that searches a key in a BST. Take a suitable binary tree and
search for a particular key in the tree by applying the discussed algorithm.
Discuss how the insert operation can be performed in a binary tree with the
help of an algorithm. Consider a binary tree and insert a particular value in
the binary tree.
Discuss how the delete operation can be performed in a binary tree. Also,
present an algorithm for the same. Consider a binary tree and delete a
particular value in the binary tree.
Discuss the process of deletion of a node in a binary tree for the following
cases by taking suitable examples:

i.
ii.
iii.
2.

3.
4.

+*-ABCD
+A+BC
/-ABC
/*A+BCD
+*AB/CD
*A+B/CD

The node is a leaf node.

The node has only one child.
The node is an internal node, i.e., has both the children.

What are heap trees? Discuss representation of a heap tree. Discuss a

minheap tree and a maxheap tree with the help of examples. What are the
operations that can be performed on a heap tree?
Discuss insertion and deletion of a node into/from a heap tree using
examples. Present an algorithm for both the operations.
Consider the BST shown in Figure 7.58: Perform the following operations
on the above BST:

i.
ii.
iii.
iv.
v.
vi.

Search a given key, say 7, in the BST.

Insert a given key, say 10, in the BST.
Insert a given key, say 52, in the BST.
Delete a given key, say 11, in the BST.
Delete a given key, say 10, in the BST.
Delete a given key, say 1, in the BST.
Display appropriate messages.

5.

6.

Fig. 7.58 A binary search tree of Q.21

2.

Consider the following list of numbers input by the user:

12, 5, 18, 90, 43, 36, 81, 10, 15, 8, 23

i.
ii.
iii.
iv.
v.
vi.
vii.
viii.

Construct a BST for the above list of non-negative

integers.
Search a given key, say 43, in the BST created.
Search a given key, say 93, in the BST created.
Insert a given key, say 48, in the BST created.
Insert a given key, say 23, in the BST created.
Delete 23 from the above BST created.
Delete 5 from the above BST created.
Delete 12 from the above BST created.
Display appropriate messages.

2.
3.

What do you mean by sorting? Discuss heap sort in detail with the help of
an example.
Consider the maxheap shown in Figure 7.59. Perform the desired
operations on the heap to obtain a sorted list of elements. Display the sorted
list obtained.

Fig. 7.59 Maxheap of Q.24

Chapter 8
Graphs
CHAPTER OUTLINE

8.1 Introduction
8.2 Graph Terminology
8.3 Representation of Graphs
8.4 Operations of Graphs
8.5 Applications of Graphs
8.1 INTRODUCTION

I want to meet Amitabh Bachchan. Do I know somebody who knows

him? Or is there a path of connected people that can lead me to this great
man? If yes, then how to represent the people and the path? The
problem becomes challenging if the people involved are spread across
various cities. For instance, my friend Mukesh from Delhi knows
Surender at Indore who in turn knows Wadegoanker residing in Pune.
Wadegoanker is a family friend of the great Sachin at Mumbai. The
friendship of Sachin and Amitabh is known world over. There is an
alternative given by Dr. Asok De. He says that his friend Basu from
Kolkata is the first cousin of Aishwarya Rai, the daughter in law of
Amitabh.
The various cities involved in this set of related people are connected by
roads as shown Figure 8.1.

Fig. 8.1 Cities connected with roads

Now the problem is how to represent the data given in Figure 8.1. Arrays
are linear by nature and every element in an array must have a unique
successor and predecessor except the first and last elements. The
arrangement shown in Figure 8.1 is not linear and, therefore, cannot be
represented by an array. Though a tree is a non-linear data structure but
it has a specific property that a node cannot have more than one parent.
The arrangement of Figure 8.1 is not following any hierarchy and,
therefore, there is no parentchild relationship.
A close look at the arrangement suggests that it is neither linear nor
hierarchical but a network of roads connecting the participating cities.
This type of structure is known as a graph. In fact, a graph is a collection
of nodes and edges. Nodes are connected by edges. Each node contains
data. For example, inFigure 8.1, the cities have been shown as nodes and
the roads as edges between them. Delhi, Bhopal, Pune, Kolkata, etc. are
nodes and the roads connecting them are edges. There are numerous
examples where the graph can be used as a data structure. Some of the
popular applications are as follows:
o Model of www: The model of world wide web (www) can be represented
by a collection of graphs (directed) wherein nodes denote the documents,
papers, articles, etc. and the edges represent the outgoing hyperlinks
between them.
o Railway system: The cities and towns of a country are connected through
railway lines. Similarly, road atlas can also be represented using graphs.
o Airlines: The cities are connected through airlines.
o Resource allocation graph: In order to detect and avoid deadlocks, the
operating system maintains a resource allocation graph for processes that
are active in the system.
o Electric circuits: The components are represented as nodes and the
wires/connections as edges.

Graph: A graph has two sets: V and E, i.e., G = <V, E>, where V is the
set of vertices or nodes and E is the set of edges or arcs. Each element of
set E is a pair of elements taken from set V. Consider Figure 8.1 and
identify the following:
V = {Delhi, Bhopal, Calcutta, Indore.}
E = {(Delhi, Kota), (Bhopal, Nagpur), (Pune, Mumbai)}
Before proceeding to discuss more on graphs, let us have a look at the
basic terminology of graphs discussed in the subsequent section.
8.2 GRAPH TERMINOLOGY

In this book, the following terms related to graphs are used:

Directed graph: A directed graph is a graph G = <V, E> with the

property that its edges have directions. An edge E: (vi, vj) means that
there is an arrow whose head is pointing to vj and the tail to vi. A
directed graph is shown in Figure 8.2(a). For example, (B, A), (A, D), (C,
D), etc are directed edges. A directed graph is also called a diagraph.
Undirected graph: An undirected graph is a graph G = <V, E> with
the property that its edges have no directions, i.e., the edges are two
ways as shown in Figure 8.2(b). For example, (B, A), (A, B), etc. are two
way edges. A graph always refers to an undirected graph.
Weighted graph: A weighted graph is a graph G = <V, E> with the
property that its edges have a number or label associated with it. The
number is called the weight of the edge. A weighted graph is shown
in Figure 8.2(c). For example, (B, A) 10 may mean that the distance of B
from A is 10m. A weighted graph can be an undirected graph or a
directed graph.
Path: It is a sequence of nodes in which all the intermediate pair nodes
between the first and the last nodes are joined by edges as shown
in Figure 8.2(d). The path from A to F is shown with the help of dark
edges. The path length is equal to the number of edges present in a path.
The path length of A-D-E-F is 3 as it contains three edges.
However in the case of weighted graphs, the path length is computed as
the sum of labels of the intermediate edges. For example, the path length
of path A-D-C-E of Figure 8.2(c) is (11 + 12 + 9), i.e., 32.

Fig. 8.2 Various kinds of graphs

It may be noted that if no node occurs more E E than once in a path then
the path is called a simple path. For example in Figure 8.2(b), A-B-C is a
simple path. However, the path A-D-E-C-D is not a simple path.
Cycle: A path that starts and ends on the same node is called a cycle. For
example, the path D- E-C-D in Figure 8.2(b) is a cycle.
Connected graph: If there exists a path between every pair of distinct
nodes, then the undirected graph is called a connected graph. For
example, the graphs shown in Figure 8.3(b), (c) and (d) are connected
graphs. However, the graph shown in Figure 8.3(a) is not a connected
graph.

Fig. 8.3 More graphs

Adjacent node: A node Y is called adjacent to X if there exists an edge
from X to Y. Y is called successor of X and X predecessor of Y. The set of
nodes adjacent to X are called neighbours of X.
Complete graph: A complete graph is a graph G = <V, E> with the
property that each node of the graph is adjacent to all other nodes of the
graph, as shown in Figure 8.3(b).
Acyclic graph: A graph without a cycle is called an acyclic graph as
shown in Figure 8.3(c). In fact, a tree is an excellent example of an
acyclic graph.

Self loop: If the starting and ending nodes of an edge are same, then
the edge is called a self loop, i.e., edge (E, E) is a self loop as shown
in Figure 8.4(a).
Parallel edges: If there are multiple edges between the same pair of
nodes in a graph, then the edges are called parallel edges. For example,
there are two parallel edges between nodes B and C of Figure 8.4(b).
Multigraphs: A graph containing self loop or parallel edges or both is
called a multigraph. The graphs shown in Figure 8.4 are multigraphs.
Simple graph: A simple graph is a graph which is free from self loops
and parallel edges.

Fig. 8.4 Multigraphs

Degree of a node or vertex: The number of edges connected to a
node is called its degree. The maximum degree of a node in a graph is
called the degree of the graph. For example, the degree of node A
in Figure 8.3(a) is 2 and that of Delhi in Figure 8.1 is 4. The degree of
graph in Figure 8.1 is 6.
However, in case of directed graphs, there are two degrees of a node
indegree and outdegree. The indgree is defined as the number of edges
incident upon the node and outdegree as the number of edges radiating
out of the node.
Consider Figure 8.3(d), the indegree and outdegree of C is 0 and 4,
respectively. Similarly, the indegree and outdegree of D is 3 and 1,
respectively.
Pendent node: A pendent node in a graph is a node whose indegree is
1 and outdegree is 0. The node F in Figure 8.3(d) is a pendent node. In
fact, in a tree all leaf nodes are necessarily pendent nodes.
8.3 REPRESENTATION OF GRAPHS

A graph can be represented in many ways and the most popular methods
are given below:

o Array-based representation
o Linked representation
o Set representation

These representations are discussed in detail in the sections that follow.

8.3.1 Array-based Representation of Graphs
A graph can be comfortably represented using an adjacency matrix
implemented through a two-dimensional array. Adjacency matrix is
discussed in detail below.
A graph of order N, i.e., having N nodes or vertices can be comfortably
modelled as a square matrix N N wherein the rows and the column
numbers represent the vertices as shown in Figure 8.5.
A cell at the cross-section of vertices can have only Boolean values, 0 or
1. An entry (vi, vj) is 1 only when there exists an edge between vi and vj,
otherwise a 0 indicates an absence of edge between the two.
The adjacency matrices for graphs of Figures 8.3(ad) are given
in Figure 8.6. From Figure 8.6(a), it may be observed that a row provides
list of successors of a node and the column provides list of predecessors.
The successors of B are A and C (see Figure 8.6(a)). An empty column
under B suggests that there are no predecessors of B.

Fig. 8.5 A matrix representing a graph

It may be noted that in the case of weighted graph (Figure 8.6 (c)), the
edge weight has been stored instead of 1 to mark the presence of an edge.
Some important properties of adjacency matrices are as follows:

Fig. 8.6 The equivalent adjacency matrices for graphs

Property 1: It may further be noted that the adjacency matrix
is symmetric for an undirected graph. This can be easily proved that an
edge (vi, vj) of an undirected graph gets an entry into an adjacency
matrix A at two locations: Aij and Aji. Thus in the matrix A for all i, j, Aij =
Aji, which is the property of a symmetric matrix. Therefore, the following
relation also holds good:
A = AT
where AT is the transpose of adjacency matrix A.
From Figure 8.6(a), it can be confirmed that in case of directed graphs,
the adjacency matrix is not symmetric and therefore the following
relation holds true.
A AT
However, for simple graphs, the diagonal elements are always 0.
Property 2: For a directed graph G with adjacency matrix A, the
diagonal of matrix A AT gives the out degrees of the corresponding
vertices.
Consider the graph and its adjacency matrix A given in Figure 8.7.

Fig. 8.7 The out degree of vertices denoted by diagonal elements of A

AT
It may be noted that the diagonal elements of A AT are {1, 2, 1, 1, 1},
which are in fact the outdegrees of vertices A, B, C, D and E, respectively.
Similarly, for a directed graph G with adjacency matrix A, the diagonal of
matrix AT A gives the in degrees of the corresponding vertices.
Consider the graph and its adjacency matrix A given in Figure 8.8.
It may be noted that the diagonal elements of AT A are {1, 0, 2, 2, 1},
which are in fact the in degrees of vertices A, B, C, D and E, respectively.

Fig. 8.8 The in degree of vertices denoted by diagonal elements of AT

A
Property 3: For a directed graph G with adjacency matrix A, the
element Aij of matrix AK (kth power of A) gives the number of paths of
length k from vertex vi to vj. This means that an element Aij of A2 will
denote the number of paths of length 2 from vertex vi to vj.
Consider Figure 8.9. A2 gives the path length of 2 from a vertex vi to vj.
For example, there is one path of length 2 from C to A (i.e., C-B-A).

Similarly, A3 gives the path length of 3 from a vertex vi to vj. For

example, there is one path of length 3 from E to D (i.e., E-C-B-D).
The major drawback of adjacency matrix is that it requires N N space
whereas most of the entries in the matrix are 0, i.e., the matrix is sparse.

Fig. 8.9 Number of paths from a vertex to another vertex

8.3.2 linked Representation of a Graph
In order to save the space, an array of linked lists is used to represent the
adjacent nodes and the data structure is called adjacency list. For
example, the adjacency list for a directed graph is shown in Figure 8.10.
Similarly, the adjacency list for an undirected graph is shown in Figure
8.11.
The above representation involves a mix of array and linked lists, i.e., it
is an array of linked lists. For pure linked representation, the array part
can be replaced by another linked list of head nodes as shown in Figure
8.12. The head node contains three parts: data, pointer to adjacent node
and a pointer to next head node as shown in Figure 8.13.

Fig. 8.10 Adjacency list for a directed graph

Fig. 8.11 Adjacency list for an undirected graph

Fig. 8.12 The linked representation of an adjacency list of a directed

graph

Fig. 8.13 The structure of a head node

Note: If in a graph G, there are N vertices and M edges, the adjacency
matrix has space complexity O(N2) and that of adjacency list is O(M).

8.3.3 Set Representation of Graphs

This representation follows the definition of the graph itself in the sense
that it maintains two sets: V- Set of Vertices and E-Set of Edges. The
set representation for various graphs is given in Figure 8.14.

Fig. 8.14 The set representation of graphs

In case of weighted graphs, the weight is also included with the edge
information making it a 3-tuple, i.e., (w, vi, vj) where w is the weight of
edge between vertices vi and vj. The set representation for a weighted
graph is given in Figure 8.15.

Fig. 8.15 The set representation of a weighted graph

This set representation is most efficient as far as storage is concerned but
is not suitable for operations that can be defined on graphs. A discussion
on operations on graphs is given in the section 8.4.
8.4 OPERATIONS OF GRAPHS

The following operations are defined on graphs:

1. Insertion: There are two major components of a graphvertex and edge.
Therefore, a node or an edge or both can be inserted into an existing graph.
2. Deletion: Similarly, a node or an edge or both can be deleted from an
existing graph.

3. Traversal: A graph may be traversed for many purposesto search a path,

to search a goal, to establish a shortest path between two given nodes, etc.

8.4.1 Insertion Operation

The insertion of a vertex and its associated edge with other vertices in an
adjacency matrix involves the following operations:
1. Add a row for the new vertex.
2. Add a column for the new vertex.
3. Make appropriate entries into the rows and columns of the adjacency
matrix for the new edges.

Consider the graph given in Figure 8.16.

Fig. 8.16 Insertion of a vertex into an undirected graph

It may be noted that a vertex called F has been added to the graph. It has
two edges (F, B) and (F, D). The adjacency matrix has been accordingly
modified to have a new row and a new column for the incoming vertex.
Appropriate entries for the edges have been made in the adjacency
matrix, i.e., the cells (F, B), (F, D), (B, F) and (D, F) have been set equal
to 1.
An algorithm for insertion of a vertex is given below:
/*Thisalgorithmusesatwodimensionalmatrix
adjMat[][]tostoretheadjacencymatrix.Anew
vertexcalledverTexisaddedtothematrixbyadding
anadditionalemptyrowandanadditionalempty
column.*/

Algorithm addVertex()
{
lastRow=lastRow+1;
lastCol=lastCol+1;
adjMat[lastRow][0]=verTex;
adjMat[0][lastCol]=verTex;
Setallelementsoflastrow=0;
SetallelementsoflastCol=0;
}
An algorithm for insertion of an edge into an undirected graph is given
below:
/*Thisalgorithmusesatwodimensionalmatrix
adjMat[][]tostoretheadjacencymatrix.Anewedge
(v1,v2)isaddedtothematrixbyaddingitsentry
(i.e.,1)intherowandcolumncorrespondingtov1
andv2,respectively.*/
Algorithm addEdge()
{
Findrowcorrespondingtov1,i.e.,rowV1;
Findcolcorrespondingtov2,i.e.,colV2;
adjMat[rowV1][ColV2]=1;/*makesymmetric
entries*/
adjMat[colV2][rowV1]=1;
}
Example 1: Write a program that implements an adjacency matrix
wherein it adds a vertex and its associated edges. The final adjacency
matrix is displayed.
Solution: Two functions
called addVertex() and addEdge() would be used for insertion of
a new vertex and its associated edges, respectively. Another function
called dispAdMat() would display the adjacency matrix.
The required program is given below:
/*Thisprogramimplementsanadjacencymatrix*/
#include<stdio.h>
#include<conio.h>
voidaddVertex(charadjMat[7][7],intnumV,char
verTex);

voidaddEdge(charadjMat[7][7],charv1,charv2,
intnumV);
voiddispAdMat(charadjMat[7][7],intnumV);
voidmain()
{/*AdjacencymatrixofFigure8.17*/
charadjMat[7][7]={_,A,B,C,D,E,,
A,0,1,0,1,0,,
B,1,0,1,0,0,,
C,0,1,0,1,1,,
D,1,0,1,0,1,,
E,0,0,1,1,0,,
,,,,,,};
intnumVertex=5;
charnewVertex,v1,v2;
charchoice;
dispAdMat(adjMat,numVertex);
printf(\nEnterthevertextobeadded);
fflush(stdin);
newVertex=getchar();
numVertex++;
addVertex(adjMat,numVertex,newVertex);
do
{
fflush(stdin);
printf(\nEnterEdge:v1v2);
scanf(%c%c,&v1,&v2);
addEdge(adjMat,v1,v2,numVertex);
dispAdMat(adjMat,numVertex);
fflush(stdin);
printf(\ndoyouwanttoaddanotheredgeY/N);
choice=getchar();
}
while((choice!=N)&&(choice!=n));
}
voidaddVertex(charadjMat[7][7],intnumV,char
verTex)
{inti;
adjMat[numV][0]=verTex;
adjMat[0][numV]=verTex;
for(i=1;i<=numV;i++)
{

adjMat[numV][i]=0;
adjMat[i][numV]=0;
}
}
i=0;
for(j=1;j<=numV;j++)
{
if(adjMat[i][j]==v1)
{
for(k=0;k<=numV;k++)
{
if(adjMat[k][0]==v2)
{
adjMat[k][j]=1;
adjMat[j][k]=1;break;/*making
symmetricentries*/
}
}
}
}
}
voiddispAdMat(charadjMat[7][7],intnumV)
{
inti,j;
printf(\nTheadjMatis\n);
for(i=0;i<=numV;i++)
{
for(j=0;j<=numV;j++)
{
printf(%c,adjMat[i][j]);
}
printf(\n);
}
printf(\n\nEnteranykeytocontinue);
getch();
}
The above program has been tested for data given in Figure 8.16. The
screenshots of the result are given in Figure 8.17.

Fig. 8.17 The adjacency matrix before and after insertion of vertex F
Note: In case of a directed graph, only one entry of the directed edge in
the adjacency matrix needs to be made. An algorithm for insertion of a
vertex in a directed graph is given below:
/*Thisalgorithmusesatwodimensionalmatrix
adjMat[][]tostoretheadjacencymatrix.Anew
vertexcalledverTexisaddedtothematrixbyadding
anadditionalemptyrowandanadditionalempty
column.*/
Algorithm addVertex()
{
lastRow=lastRow+1;
lastCol=lastCol+1;
adjMat[lastRow][0]=verTex;
adjMat[0][lastCol]=verTex;
Setallelementsoflastrow=0;
SetallelementsoflastCol=0;
}
An algorithm for insertion of an edge into a directed graph is given
below:
/*Thisalgorithmusesatwodimensionalmatrix
adjMat[][]tostoretheadjacencymatrix.Anewedge

(v1,v2)isaddedtothematrixbyaddingitsentry
(i.e.,1)intherowcorrespondingtov1.*/
Algorithm addEdge()
{
Findrowcorrespondingtov1,i.e.,rowV1;
Findcolcorrespondingtov2,i.e.,colV2;
adjMat[rowV1][ColV2]=1;
}
Accordingly, the function addEdge() needs to modified so that it
makes the adjacency matrix assymmetric. The modified function is given
below:
/*Themodifiedfunctionforadditionofedgesin
directedgraphs*/
voidaddEdge(charadjMat[7][7],charv1,charv2,
intnumV)
{inti,j,k;
j=0;
for(i=1;i<=numV;i++)
{
if(adjMat[i][j]==v1)
{
for(k=1;k<=numV;k++)
{
if(adjMat[0][k]==v2)
{
adjMat[i][k]=1;break;
}
}
}
}
}
8.4.2 Deletion Operation
The deletion of a vertex and its associated edges from an adjacency
matrix, for directed and undirected graph, involves the following
operations:
1. Deleting the row corresponding to the vertex.
2. Deleting the col corresponding to the vertex. The algorithm is straight
forward and given below:

Algorithm delVertex(verTex)
{
findtherowcorrespondingtoverTexandsetallits
elements=0;
findthecolcorrespondingtoverTexandsetallits
elements=0;
}
The deletion of an edge from a graph requires different treatment for
undirected and directed graphs. Both the cases of deletion of edges are
given below:
An algorithm for deletion of an edge from an undirected graph is given
below:
/*Thisalgorithmusesatwodimensionalmatrix
adjMat[][]tostoretheadjacencymatrix.Anewedge
(v1,v2)isdeletedfromthematrixbydeletingits
entry(i.e.,0)fromtherowandcolumn
correspondingtov1andv2,respectively.*/
Algorithm delEdge()/*undirectedgraph*/
{
Findrowcorrespondingtov1,i.e.,rowV1;
Findcolcorrespondingtov2,i.e.,colV2;
adjMat[rowV1][ColV2]=0;/*makesymmetricentries
*/
adjMat[colV2][rowV1]=0;
}
An algorithm for deletion of an edge from a directed graph is given
below:
/*Thealgorithmusesatwodimensionalmatrix
adjMat[][]tostoretheadjacencymatrix.Anedge
(v1,v2)isdeletedfromthematrixbydeletingits
entry(i.e.,0)fromtherowcorrespondingtov1*/
Algorithm delEdge()/*directedgraph*/
{
Findrowcorrespondingtov1,i.e.,rowV1;
Findcolcorrespondingtov2,i.e.,colV2;
adjMat[rowV1][ColV2]=0;
}

Example 2: Write a program that deletes a vertex and its associated

edges from an adjacency matrix. The final adjacency matrix is displayed.
Solution: A function called delVertex() would be used for deletion
of a vertex and its associated edges from an adjacency matrix. Another
function called dispAdMat() would display the adjacency matrix.
The required program is given below:
/*Thisprogramdeletesavertexfromanadjacency
list*/
#include<stdio.h>
#include<conio.h>
voiddelVertex(charadjMat[7][7],intnumV,char
verTex);
voiddispAdMat(charadjMat[7][7],intnumV);
voidmain()
{
charadjMat[7][7]={,A,B,C,D,E,,
A,0,1,0,1,0,,
B,1,0,1,0,0,,
C,0,1,0,1,1,,
D,1,0,1,0,1,,
E,0,0,1,1,0,,
,,,,,,};
intnumVertex=5;
charVertex;
dispAdMat(adjMat,numVertex);
printf(\nEnterthevertextobedeleted);
fflush(stdin);
Vertex=getchar();
delVertex(adjMat,numVertex,Vertex);
dispAdMat(adjMat,numVertex);
}
voiddelVertex(charadjMat[7][7],intnumV,char
verTex)
{inti,j,k;
j=0;
for(i=1;i<=numV;i++)
{
if(adjMat[i][j]==verTex)
{

adjMat[i][0]=_;
for(k=1;k<=numV;k++)
{
adjMat[i][k]=0;
}
}
}
i=0;
for(j=1;j<=numV;j++)
{
if(adjMat[i][j]==verTex)
{
adjMat[0][j]=_;
for(k=1;k<=numV;k++)
{
adjMat[k][j]=0;
}
}
}
}
voiddispAdMat(charadjMat[7][7],intnumV)
{
inti,j;
printf(\nTheadjMatis\n);
for(i=0;i<=numV;i++)
{
for(j=0;j<=numV;j++)
{
printf(%c,adjMat[i][j]);
}
printf(\n);
}
printf(\n\nEnteranykeytocontinue);
getch();
}
The screenshots of the output of the program are shown in Figure 8.18.

Fig. 8.18 The adjacency matrix before and after deletion of vertex B
Example 3: Write a program that deletes an edge from an adjacency
matrix of an undirected graph. The final adjacency matrix is displayed.
Solution: A function called delEdge() would be used for deletion of
an edge from an adjacency matrix. Another function called dispAdMat(
) would display the adjacency matrix.
The required program is given below:
/*Thisprogramdeletesanedgeofanundirected
graphfromanadjacencymatrix*/
#include<stdio.h>
#include<conio.h>
voiddelEdge(charadjMat[7][7],charv1,charv2,
intnumV);
voiddispAdMat(charadjMat[7][7],intnumV);
voidmain()
{
charadjMat[7][7]={,A,B,C,D,E,,
A,0,1,0,1,0,,
B,1,0,1,0,0,,
C,0,1,0,1,1,,
D,1,0,1,0,1,,
E,0,0,1,1,0,,

,,,,,,};
intnumVertex=5;
charv1,v2;
dispAdMat(adjMat,numVertex);
printf(\nEntertheedgetobedeleted);
fflush(stdin);
printf(\nEnterEdge:v1v2);
scanf(%c%c,&v1,&v2);
delEdge(adjMat,v1,v2,numVertex);
dispAdMat(adjMat,numVertex);
}
voiddelEdge(charadjMat[7][7],charv1,charv2,
intnumV)
{inti,j,k;
i=0;
for(j=1;j<=numV;j++)
{
if(adjMat[i][j]==v1)
{
for(k=0;k<=numV;k++)
{
if(adjMat[k][0]==v2)
{
adjMat[k][j]=0;
adjMat[j][k]=0;break;/*making
symmetricentries*/
}
}
}
}
}
voiddispAdMat(charadjMat[7][7],intnumV)
{
inti,j;
printf(\nTheadjMatis\n);
for(i=0;i<=numV;i++)
{
for(j=0;j<=numV;j++)
{
printf(%c,adjMat[i][j]);
}

printf(\n);
}
printf(\n\nEnteranykeytocontinue);
getch();
}
The screenshots of the output of the program are given in Figure 8.19.

Fig. 8.19 The adjacency matrix before and after deletion of edge (D, A)
The function delEdge() needs to modified for directed graphs so
that it does not make the adjacency matrix as symmetric. The modified
function is given below:
/*Themodifiedfunctionfordeletionofedgesof
directedgraphs*/
voiddelEdge(charadjMat[7][7],charv1,charv2,
intnumV)
{inti,j,k;
j=0;
for(i=1;i<=numV;i++)
{
if(adjMat[i][j]==v1)
{
for(k=1;k<=numV;k++)
{
if(adjMat[i][k]==v2)
{

adjMat[i][k]=0;break;
}
}
}
}
}
8.4.3 Traversal of a Graph
Travelling a graph means that one visits the vertices of a graph at least
once. The purpose of the travel depends upon the information stored in
the graph. For example on www, one may be interested to search a
particular document on the Internet while on a resource allocation graph
the operating system may search for deadlocked processes.
In fact, search has been used as an instrument to find solutions. The epic
Ramayana uses the search of goddess Sita to find and settle the
problem of Demons of that yuga including Ravana. Nowadays, the
search of graphs and trees is very widely used in artificial intelligence.
A graph can be traversed in many ways but there are two popular
methods which are very widely used for searching graphs; they are:
Depth First Search (DFS) and Breadth First Search (BFS). A detailed
discussion on both is given in the subsequent sections.
8.4.3.1 Depth First Search (DFS)
In this method, the travel starts from a vertex then carries on to its
successors, i.e., follows its outgoing edges. Within successors, it travels
their successor and so on. Thus, this travel is same as inorder travel of a
tree and, therefore, the search goes deeper into the search space till no
successor is found. Once the search space of a vertex is exhausted, the
travel for next vertex starts. This carries on till all the vertices have been
visited. The only problem is that the travel may end up in a cycle. For
example, in a graph, there is a possibility that the successor of a
successor may be the vertex itself and the situation may end up in an
endless travel, i.e., a cycle.
In order to avoid cycles, we may maintain two
queues: notVisited and Visited. The vertices that have already been
visited would be placed on Visited. When the successors of a vertex are
generated, they are placed on notVisited only when they are not
already present on both Visited and notVisitedqueues thereby
reducing the possibility of a cycle.
Consider the graph shown in Figure 8.20. The trace of DFS travel of the
given graph is given in Figures 8.20 and 8.21, through entries into the
two queues Visited and notVisited.

Fig. 8.20 The trace of DFS travel of a graph

The algorithm for DFS travel is given below:
Algorithm DFS(firstVertex)
{
addfirstVertexonnotVisited;
placeNULLonVisited;/*initiallyVisitedis
empty*/
while(notVisited!=NULL)
{
removevertexfromnotVisited;
generatesuccessorsofvertex;
foreachsuccessorofvertexdo
{

if(successorisnotpresentonVisitedAND
successorisnotpresentonnotVisited)Thenadd
vertexonfrontofnotVisited;
}
if(vertexisnotpresentonVisited)Thenaddvertex
onVisited;
}
}
The final visited queue contains the nodes visited during DFS travel, i.e.,
A B C F E G D.

Fig. 8.21 The trace of DFS travel of a graph

Example 4: Write a program that travels a given graph using DFS
strategy.

Solution: We would use the following functions:

1. voidaddNotVisited(charQ[],int*first,char
vertex);:
ThisfunctionaddsavertexonnotVisitedqueueinfront
ofthequeue.
2. voidaddVisited(charQ[],int*last,charvertex);
ThisfunctionaddsavertexonVisitedqueue.
3. charremoveNode(charQ[],int*first);
ThisfunctionremovesavertexfromnotVisitedqueue.
4. intfindPos(charadjMat[8][8],charvertex);
Thisfunctionfindsthepositionofavertexin
adjacencymatrix,i.e.,adjMat.
5. voiddispAdMat(charadjMat[8][8],intnumV);
Thisfunctiondisplaysthecontentsofadjacencymatrix,
i.e.,adjMat.
6. intifPresent(charQ[],intlast,charvertex);
Thisfunctioncheckswhetheravertexispresentona
queueornot.
7. voiddispVisited(charQ[],intlast);
ThisfunctiondisplaysthecontentsofVisitedqueue.

We would use the adjacency matrix of graph shown in Figures

8.20 and 8.21. The required program is given below:
/*ThisprogramtravelsagraphusingDFSstrategy*/
#include<stdio.h>
#include<conio.h>
voidaddNotVisited(charQ[],int*first,char
vertex);
voidaddVisited(charQ[],int*last,charvertex);
charremoveNode(charQ[],int*first);
intfindPos(charadjMat[8][8],charvertex);
voiddispAdMat(charadjMat[8][8],intnumV);
intifPresent(charQ[],intlast,charvertex);
voiddispVisited(charQ[],intlast);
voidmain()
{inti,j,first,last,pos;
charadjMat[8][8]={,A,B,C,D,
E,F,G,
A,0,1,0,1,0,0,0,
B,1,0,1,0,0,1,1,
C,0,1,0,1,0,1,1,

D,1,0,1,0,1,1,0,
E,0,0,0,1,0,1,0,
F,0,1,1,1,1,0,0,
G,0,1,1,0,0,0,0};
intnumVertex=7;
charvisited[7],notVisited[7];
charvertex,v1,v2;
clrscr();
dispAdMat(adjMat,numVertex);
first=7;
last=1;
addNotVisited(notVisited,&first,adjMat[1][0]);
while(first<7)
{
vertex=removeNode(notVisited,&first);
if(!ifPresent(visited,last,vertex))
addVisited(visited,&last,vertex);
pos=findPos(adjMat,vertex);
for(i=7;i>=1;i)
{
if(adjMat[pos][i]==1)
{
if(!ifPresent(notVisited,7,adjMat[0][i])&&
(!ifPresent(visited,7,adjMat[0][i])))
addNotVisited(notVisited,&first,adjMat[0]
[i]);
}
}
}
printf(\nFinalvisitednodes...);
dispVisited(visited,last);
}
voiddispAdMat(charadjMat[8][8],intnumV)
{
inti,j;
printf(\nTheadjMatis\n);
for(i=0;i<=numV;i++)
{
for(j=0;j<=numV;j++)
{
printf(%c,adjMat[i][j]);

}
printf(\n);
}
printf(\n\nEnteranykeytocontinue);
getch();
}
/*Thisfunctioncheckswhetheravertexispresent
onaQornot*/
intifPresent(charQ[],intlast,charvertex)
{
inti,result=0;
for(i=0;i<=last;i++)
{
if(Q[i]==vertex)
{result=1;
break;}
}
returnresult;
}
voidaddVisited(charQ[],int*last,charvertex)
{
*last=*last+1;
Q[*last]=vertex;
}
voidaddNotVisited(charQ[],int*first,char
vertex)
{
*first=*first1;
Q[*first]=vertex;
}
voiddispVisited(charQ[],intlast)
{
inti;
printf(\nThevisitednodesare);
for(i=0;i<=last;i++)
printf(%c,Q[i]);
}
charremoveNode(charQ[],int*first)
{
charch;

ch=Q[*first];
Q[*first]=#;
*first=*first+1;
returnch;
}
intfindPos(charadjMat[8][8],charvertex)
{
inti;
for(i=0;i<=7;i++)
{
if(adjMat[i][0]==vertex)
returni;
}
return0;
}
The output of the program is shown in Figure 8.22. The final list of
visited nodes matches with the results obtained in Figures 8.20 and 8.21,
i.e., the visited nodes are A B C F G D E.

Fig. 8.22 The output of DFS travel

8.4.3.2 Breadth First Search (BFS)
In this method, the travel starts from a vertex then carries on to its
adjacent vertices, i.e., follows the next vertex on the same level. Within a
level, it travels all its siblings and then moves to the next level. Once the
search space of a level is exhausted, the travel for the next level starts.
This carries on till all the vertices have been visited. The only problem is
that the travel may end up in a cycle. For example, in a graph, there is a
possibility that the adjacent of an adjacent may be the vertex itself and
the situation may end up in an endless travel, i.e., a cycle.
In order to avoid cycles, we may maintain two queues
notVisited and Visited. The vertices that have already been visited

would be placed on Visited. When the successor of a vertex are

generated, they are placed on notVisited only when they are not
already present on both VisitedandnotVisited queues, thereby
reducing the possibility of a cycle.
Consider the graph given in Figure 8.23. The trace of BFS travel of the
given graph is given in Figures 8.23 and 8.24, through entries into the
two queuesVisited and notVisited.
The Visited queue contains the nodes visited during BFS travel, i.e., A B
D C F G E.
The algorithm for BFS travel is given below:
Algorithm BFS(firstVertex)
{
addfirstVertexonnotVisited;
placeNULLonVisited;/*initiallyVisitedisempty
*/
while(notVisited!=NULL)
{
removevertexfromnotVisited;
generateadjacentsofvertex;
foreachadjacentofvertexdo
{
if(adjacentisnotpresentonVisitedAND
adjacentisnotpresentonnotVisited)Thenadd
adjacentatRearonnotVisited;
}
if(vertexisnotpresentonVisited)Thenadd
vertexonVisited;
}
}

Fig. 8.23 The trace of BFS travel of a graph

Example 5: Write a program that travels a given graph using breadth
first search strategy.
Solution: We would use the following functions:

1. voidaddNotVisited(charQ[],int*first,char
vertex);:
ThisfunctionaddsavertexonnotVisitedqueueatRear
ofthequeue.
2. voidaddVisited(charQ[],int*last,charvertex);
3. charremoveNode(charQ[],int*first);
4. intfindPos(charadjMat[8][8],charvertex);
5. voiddispAdMat(charadjMat[8][8],intnumV);

6. intifPresent(charQ[],intlast,charvertex);
7. voiddispVisited(charQ[],intlast);

The purpose of the above functions has been explained in Example4.

The functionaddNotVisited() has been slightly modified.

Fig. 8.24 The trace of BFS travel of a graph

The adjacency matrix of graph shown in Figures 8.20 and 8.21 is used.
The program of Example 4 has been suitably modified which is given
below:
/*ThisprogramtravelsagraphusingBFSstrategy*/
#include<stdio.h>
#include<conio.h>
voidaddNotVisited(charQ[],int*first,char
vertex);

voidaddVisited(charQ[],int*last,charvertex);
charremoveNode(charQ[],int*first);
intfindPos(charadjMat[8][8],charvertex);
voiddispAdMat(charadjMat[8][8],intnumV);
intifPresent(charQ[],intlast,charvertex);
voiddispVisited(charQ[],intlast);
voidmain()
{inti,j,first,last,pos,Rear;
charadjMat[8][8]={,A,B,C,D,
E,F,G,
A,0,1,0,1,0,0,0,
B,1,0,1,0,0,1,1,
C,0,1,0,1,0,1,1,
D,1,0,1,0,1,1,0,
E,0,0,0,1,0,1,0,
F,0,1,1,1,1,0,0,
G,0,1,1,0,0,0,0};
intnumVertex=7;
charvisited[7],notVisited[7];
charvertex,v1,v2;
clrscr();
dispAdMat(adjMat,numVertex);
first=1;
Rear=1;
last=1;
addNotVisited(notVisited,&Rear,adjMat[1][0]);
while(first<=Rear)
{
vertex=removeNode(notVisited,&first);
if(!ifPresent(visited,last,vertex))
addVisited(visited,&last,vertex);
pos=findPos(adjMat,vertex);
for(i=1;i<=7;i++)
{
if(adjMat[pos][i]==1)
{
if(!ifPresent(notVisited,7,adjMat[0][i])&&
(!ifPresent(visited,7,adjMat[0][i])))
{
addNotVisited(notVisited,&Rear,adjMat[0]
[i]);

}
}
}
}
printf(,\nFinalvisitednodes,,);
dispVisited(visited,last);
}
voiddispAdMat(charadjMat[8][8],intnumV)
{
inti,j;
printf(,\nTheadjMatis\n,);
for(i=0;i,<=numV;i++)
{
for(j=0;j,<=numV;j++)
{
printf(,%c,,adjMat[i][j]);
}
printf(,\n,);
}
printf(,\n\nEnteranykeytocontinue,);
getch();
}
/*Thisfunctioncheckswhetheravertexispresent
onaQornot*/
intifPresent(charQ[],intlast,charvertex)
{
inti,result=0;
for(i=0;i,<=last;i++)
{
if(Q[i]==vertex)
{result=1;
break;}
}
returnresult;
}
voidaddVisited(charQ[],int*last,charvertex)
{
*last=*last+1;
Q[*last]=vertex;
}
voidaddNotVisited(charQ[],int*Rear,charvertex)

{
*Rear=*Rear+1;
Q[*Rear]=vertex;
}
voiddispVisited(charQ[],intlast)
{
inti;
printf(\nThevisitednodesare);
for(i=0;i<=last;i++)
{
printf(%c,Q[i]);
}
}
charremoveNode(charQ[],int*first)
{
charch;
ch=Q[*first];
Q[*first]=#;
*first=*first+1;
returnch;
}
intfindPos(charadjMat[8][8],charvertex)
{
inti;
for(i=0;i<=7;i++)
{
if(adjMat[i][0]==vertex)
returni;
}
return0;
}
The output of the program is shown in Figure 8.25. The final list of
visited nodes matches with the results obtained in Figures 8.23 and 8.24,
i.e., the visited nodes are A B D C F G E.

Fig. 8.25 The output of BFS travel of Example 5

Note: Implementation of operations on graphs using adjacency matrix
is simple as compared to adjacency list. In fact, adjacency list would
require the extra overhead of maintaining pointers. Moreover, the linked
lists connected to vertices are independent and it would be difficult to
establish cross-relationship between the vertices contained on different
lists, which is otherwise possible by following a column in adjacency
matrix.
8.4.4 Spanning Trees
A connected undirected graph has the following two properties:
o There exists a path from every node to every other node.
o The edges have no associated directions.

For example, the graph shown in Figure 8.26 is a connected undirected

graph.
We know that a tree is a special graph which does not contain a cycle.
Thus, we can remove some edges from the graph of Figure 8.26 such that
it is still connected but has no cycles. The various subgraphs or trees so
obtained are shown in Figure 8.27.

Fig. 8.26 A connected undirected graph

Fig. 8.27 The acyclic sub-graphs or trees

Each tree shown in Figure 8.27 is called a spanning tree of the graph
given in Figure 8.26. Precisely, we can define a spanning tree of a
graph G as a connected sub-graph, containing all vertices of G,
with no cycles. It may be notedthat there will be exactly n-1 edges in a
spanning tree of a graph G having n vertices. Since a spanning tree has
very less number of edges, it is also called as a sparse graph.

Fig. 8.28 DFS and BFS spanning trees

A graph can be traveled in two ways: Breadth First Search (BFS) and
Depth First Search (DFS). In these travels, the cycles are avoided.
Therefore the travels result in two trees: BFS-spanning tree and DFSspanning trees. The DFS and BFS spanning trees corresponding to graph
of Figure 8.26 are given in Figure 8.28
8.4.4.1 Minimum Cost Spanning
Trees From previous section, we know that multiple spanning trees can
be drawn from a connected undirected graph. If it is weighted graph then
the cost of each spanning tree can be computed by adding the weights of
all the edges of the spanning tree. If the weights of the edges are unique
then there will be one spanning tree whose cost is minimum of all. The
spanning tree with minimum cost is called as minimum spanning tree or
Minimum-cost Spanning Tree (MST).
Having introduced the spanning trees, lets us now see where they can be
applied. An excellent situation that comes to mind is a housing colony
wherein the welfare association is interested to lay down the electric
cables, telephone lines and water pipelines. In order to optimize, it
would be advisable to connect all the houses with minimum number of

connecting lines. A critical look at the situation indicates that a spanning

tree would be an ideal tool to model the connectivity layout. In case the
houses are not symmetrical or are not uniformly distanced then an MST
can be used to model the connectivity layout. Computer network cabling
layout is another example where an MST can be usefully applied.
There are many algorithms available for finding an MST. However,
following two algorithms are very popular.
o Kruskals algorithm
o Prims algorithm

A detailed discussion on these algorithms is given below:

Kruskals algorithm: In this algorithm, the edges of graph of n nodes
are placed in a list in ascending order of their weights. An empty
spanning tree T is taken. The smallest edge is removed from the list and
added to the tree only if it does not create a cycle. The process is repeated
till the number edges is less than or equal to n-1 or the list of edges
becomes empty. In case the list becomes empty, then it is inferred that
no spanning tree is possible.
Consider the graph given in Figure 8.29
The sorted list of edges in increasing order of their weight is given below:

Fig. 8.29 A connected undirected graph

Edge

Weight

AF

10

CD

12

BG

14

BC

16

DG

18

DE

22

EG

24

EF

25

AB

26

Now add the edges starting from shortest edge in sequence from A-F to
A-B. The step by step creation of the spanning tree is given in Figure
8.30 (Steps 1-6)
Having obtained an MST, we are now in a position to appreciate the
algorithm proposed by Kruskal as given below:
Input:

T: An empty spanning tree

E: Set of edges
N: Number of nodes in a connected undirected graph

Fig. 8.30 Creation of spanning tree

Kruskals algorithm MST()
{
Step1.sorttheedgescontainedinEinascending
orderoftheirweighs
2.count=0
3.pickanedgeoflowestweightfromE.Deleteit
fromE
4.addtheedgeintoTifitdoesnotcreatea
cycleelsediscarditandrepeatstep3
5.count=count+1
6.if(count<N1)&&Enotemptyrepeatsteps2
5
7.ifTcontainslessthann1edgesandEis
emptyreportnospanningtreeispossible.Stop.
8.returnT.
)

Example 6: Find the minimum spanning tree for the graph given
in Figure 8.31 using Kruskals algorithm.

Fig. 8.31 A connected undirected graph

Solution: The sorted list of edges in increasing order of their weight is
given below:
Edge

Weight

FG

BC

AG

EF

AF

CG

AB

CD

BG

EG

DG

DE

Now add the edges starting from shortest edge in sequence from F-G to
D-E. The step by step creation of the spanning tree is given in Figure
8.32 (Steps 1-6)
Prims algorithm
In this algorithm the spanning tree is grown in stages. A vertex is chosen
at random and included as the first vertex of the tree. Thereafter a vertex
from remaining vertices is so chosen that it has a smallest edge to a
vertex present on the tree. The selected vertex is added to the tree. This
process of addition of vertices is repeated till all vertices are included in
the tree. Thus at given moment, there are two set of vertices: Ta set of
vertices already included in the tree, and Eanother set of remaining
vertices. The algorithm for this procedure is given below:
Input:

T: An empty spanning tree

E: Empty set of edges
N: Number of nodes in a connected undirected graph

Fig. 8.32 Creation of spanning tree

AlgorithmPrim()
{

Step1.TakerandomnodevaddittoT.Addadjacent
nodesofvtoE
2.while(numberofnodesinT<N)
{2.1ifEisemptythenreportnospanningtree
ispossible.
Stop.
2.3foranodexofTchoosenodeysuchthat
theedge(x,y)haslowestcost
2.4ifnodeyisinTthendiscardtheedge
(x,y)andrepeatstep2
2.5addthenodeyandtheedge(x,y)toT.
2.6delete(x,y)fromE.
2.7addtheadjacentnodesofytoE.
}
3.returnT
}
It may be noted that the implementation of this algorithm would require
the adjacency matrix representation of a graph whose MST is to be
generated.
Let us consider the graph of Figure 8.31. The adjacency matrix of the
graph is given in Figure 8.33
Let us now construct the MST from the adjacency matrix given in Figure
8.33. Let us pick A as the starting vertex. The nearest neighbour to A is
found by searching the smallest entry (other than 0) in its row. Since 3 is
the smallest entry, G becomes its nearest neighbour. Add G to the sub
graph as edge AG (see Figure 8.33). The nearest neighbour to sub graph
A-G is F as it has the smallest entry i.e. 1. Since by adding F, no cycle is
formed then it is added to the sub-graph. Now closest to A-G-F is the
vertex E. It is added to the subgraph as it does form a cycle. Similarly
edges G-C, C-B, and C-D are added to obtain the complete MST. The step
by step creation of MST is shown in Figure 8.33.

Fig. 8.33 The adjacency matrix and the different steps of obtaining MST
8.4.5 Shortest path problem
The communication and transport networks are represented using
graphs. Most of the time one is interested to traverse a network from a
source to destination by a shortest path. Foe example, how to travel from
Delhi to Puttaparthy in Andhra Pradesh? What are the possible routes
and which one is the shortest? Similarly, through which route to send a
data packet from a node across a computer network to a destination
node? Many algorithms have been proposed in this regard but the most
important algorithms are Warshalls algorithm, Floyds algorithm, and
Dijkstras algorithm.
8.4.5.1 Warshalls Algorithm
This algorithm computes a path matrix from an adjacency matrix A. The
path matrix has the following property:

Note: The path can be of length one or more

The path matrix is also called the transitive closure of A i.e. if path [i][k]
and Path [k][j] exist then path [1][j] also exists.
We have already introduced this concept in Section 8.3.1 vide Property 3
wherein it was stated thatFor a given adjacency matrix A, the element
Aij of matrix Ak (kth power of A) gives the number of paths of length k
from vertex vi to vj. Therefore, the transitive closure of an adjacency
matrix A is given by the following expression:
Path = A + A1 + A2 + + An

Fig. 8.34
The above method is compute intensive as it requires a series of matrix
multiplication operations to obtain a path matrix.
Example 7: Compute the transitive closure of the graph given in Figure
8.34.
Solution: The step wise computation of the transitive closure of the
graph is given in Figure 8.35.

Fig. 8.35 Transitive closure of a graph

Warshall gave a simpler method to obtain a path or reachability matrix
of a directed graph (diagraph) with the help of only AND or OR Boolean
operators. For example, it determines whether a path form vertex vi to vj
through vk exists or not by the following formula
Path[i][j]=Path[i][j]V([i][k]Path[k][j])
Thus, Warshalls algorithm also computes the transitive closure of an
adjacency matrix of a diagraph.
The algorithm is given below:
Input:AdjacencymatrixA[][]ofordern*n
PathMatrixPath[][]ofordern*n,initiallyempty
AlgorithmWarshall()
{/*copyAtoPath*/
for(i=1;i<=n;i++)

{
for(j=1;j<=n;j++)
{
Path[i][j]=A[i][j];
}
}
/*Findpathfromvitovjthroughvk*/
for(k=1;k<=n;k++)
{for(i=1;i<=n;i++)
{
for(j=1;j<=n;j++)
{Path[i][j]=Path[i][j]||(Path[i][k]&&
Path[k][j]);
}
}
}
returnPath;
}
Example 8: Apply Warshalls on the graph given in Figure 8.34 to
obtain its path matrix
Solution: The step by step of the trace of Warshalls algorithm is given
in Figure 8.36.

Fig. 8.36
8.4.5.2 Floyds Algorithm
The Floyds algorithm is almost same as Warshalls algorithm. The
Warshals algorithm establishes as to whether a path between vertices vi
and vj exists or not? The Floyds algorithm finds a shortest path between
every pair of vertices of a weighted graph. It uses the following data
structures:
1. A cost matrix, Cost [][] as defined below:

2. A distance matrix Dist[][] wherein an entry D[i][j] is the distance of the

shortest path from vertex vi to vj.

In fact, for given vertices vi and vj, it finds out a path from vi to vj
through vk. If vi-vk-vj path is shorter than the direct path from vi to vj
then vi-vk-vj is selected. It uses the following formula (similar to
Warshall):
Dist [i][j] = min ( Dist[i][j], ( Dist[i][k] + Dist [k][j])
The algorithm is given below:
AlgorithmFloyd()
{
for(i=1;i<=n;i++)
for(j=1;j<=n;j++)
Dist[i][j]=Cost[i][j];/*copycostmatrixto
distancematrix*/
for(k=1;k<=n;k++)
{for(i=1;i<=n;i++)
{for(j=1;j<=n;j++)
Dist[i][j]=min(Dist[i][j],(Dist[i][k]+
Dist[k][j];
}
}
}
Example 9: For the weighted graph given in Figure 8.37, find all pairs
shortest paths using Floyds algorithm.
Solution: The cost matrix and the step by step of the trace of Floyds
algorithm is given in Figure 8.38.

Fig. 8.37
8.4.5.3 Dijkstras Algorithm
This algorithm finds the shortest paths from a certain vertex to all
vertices in the network. It is also known as single source shortest path
problem. For a given weighted graph G = (V, E), one of the vertices v0 is
designated as a source. The source vertex v0 is placed on a set S with its
shortest path taken as zero. Where the set S contains those vertices
whose shortest paths are known. Now iteratively, a vertex (say vi) from
remaining vertices is added to the set S such that its path is shortest from
the source. In fact this shortest path from v0 to vi passes only through
the vertices present in set S. The process stops as soon as all the vertices
are added to the set S.

Fig. 8.38
The dijkstra algorithm uses following data structures:
1. A vector called Visited[]. It contains the vertices which have been
visited. The vertices are numbered from 1 to N. Where N is number of
vertices in the graph. Initially Visited is initialized as zeros. As soon as a
vertex i is visited (i.e. its shortest distance is found) Visited[i] is set to 1.
2. A two dimensional matrix Cost[][]. An entry Cost[i][j] contains the
cost of an edge (i, j). If there is no edge between vertices i and j then cost[i]
[j] contains a very large quantity (say 9999).
3. A vector dist[]. An entry dist[i] contains the shortest distance of ith vertex
from the starting vertex v1.
4. A vector S contains those vertices whose shortest paths are now known.

The algorithm is given below:

Algorithmdijkstra()
{
S[1]=1;Visited[1]=1;dist[1]=0;/*Initialize.
Placev1onS*/
for(i=2;i<=N;i++;)
{
Visited[i]=0;
dist[i]=cost[1][i]
}
for(i=2;i<=N;i++;)
/*Findvertexwithminimumcostanditsnumber
storedinpos*/
{
min=9999;pos=1;
for(j=2;j<=N;j++)
{
if(visited[j]=0&&dist[j]<min)
{
pos=j;
min=dist[j];
}
}
/*Addthevertex(i.e.pos)
visited[pos]=1;S[i]=pos;
/*foreachvertexadjacenttopos,updatedistances
*/
for(k=2;k<=N;k++)
{
if(visited[k]==0)
if(dist(pos)+cost[pos][k]<dist[k])
dist[k]=dist(pos)+cost[pos][k];
}
}
Consider the graph given in Figure 8.39.
The corresponding cost matrix and the stepwise construction of shortest
cost of all vertices with respect to vertex V1 is given in Figure 8.40.

Fig. 8.39 An undirected weighted graph

Fig. 8.40 Cost matrix and different steps of Dijkstras algorithm

Fig. 8.41
Example 10: For the graph given in Figure 8.41 , find the shortest path
from vertex A to all the remaining verices.
Solution: Let us number the vertices A-G as 1-7. The corresponding
cost matrix and the stepwise construction of shortest cost of all vertices
with respect to vertex 1 (i.e. A) is given in Figure 8.42.

Fig. 8.42 Cost matrix and different steps of Dijkstras algorithm

8.5 APPLICATIONS OF GRAPHS

There are numerous applications of graphs. Some of the important

applications are as follows:
1. Model of www: The model of world wide web (www) can be represented
by a graph (directed) wherein nodes denote the documents, papers, articles,
etc. and the edges represent the outgoing hyperlinks between them as
shown in Figure 8.43.

Fig. 8.43 The graphic representation of www

It may be noted that a page on site S1 has a link to page on site S4. The
pages on S4 refer to pages on S2 and S3, and so on. This hyperlink structure
is the basis of the web structure similar to the web created by a spider and,
hence, the name world wide web.
2. Resource allocation graph: In order to detect and avoid deadlocks, the
operating system maintains a resource allocation graph for processes that
are active in the system. In this graph, the processes are represented as
rectangles and resources as circles. Number of small circles within a
resource node indicates the number of instances of that resource available
to the system. An edge from a process P to a resource R, denoted by (P, R),
is called a request edge. An edge from a resource R to a process P, denoted
by (R, P), is called an allocation edge as shown in Figure 8.44.
It may be noted from the resource allocation graph of Figure 8.27 that the
process P1 has been allocated one instance of resource R1 and is requesting
for an instance of resource R2. Similarly, P2 is holding an instance of both
R1 and R2. P3 is holding an instance of R2 and is asking for an instance of
R1. The edge (R1, P1) is an allocation edge whereas the edge (P1, R2) is a
request edge.
An operating system maintains the resource allocation graph of this kind
and monitors it from time to time for detection and avoidance of deadlock
situations in the system.
3. Colouring of maps: Colouring of maps is an interesting problem wherein
it is desired that a map has to be coloured in such a fashion that no two
adjacent countries or regions have the same colour. The constraint is to use
minimum number of colours. A map (see Figure 8.45) can be represented as
a graph wherein a node represents a region and an edge between two
regions denote that the two regions are adjacent.

Fig. 8.44 resource allocation graph

4. Scene graphs: The contents of a visual scene are also managed by using
graph data structure. Virtual reality modelling language (VRML) supports

scene graph programming model. This model is used by MPEG-4 to

represent multimedia scene composition.

Fig. 8.45 The graphic representation of a map

A scene graph comprises nodes and edges wherein a node represents
objects of the scene and the edges define the relationship between the
objects. The root node called parent is the entry point. Every other node
in the scene graph has a parent leading to an acyclic graph or a tree like
structure as shown inFigure 8.46.

Fig. 8.46 A scene graph for a visual scene

For more reading on scene graphs, the reader may refer to the paper
Understanding Scene Graphs by Aeron E. Walsh, chairman of Mantis
development at Boston College.
Besides above, the other popular applications of graphs cited in books
are:
o
o
o
o

Shortest path problem

Spanning trees
Binary decision graph
Topological sorting of a graph

EXERCISES
1. Explain the following graph theory terms:

a.
b.
c.
d.
e.
f.

Node
Edge
Directed graph
Undirected graph
Connected graph
Disconnected graph

2. What is a graph and how is it useful?

3. Draw a directed graph with five vertices and seven edges. Exactly one of the
edges should be a loop, and should not have any multiple edges.
4. Draw an undirected graph with five edges and four vertices. The vertices
should be called v1, v2, v3 and v4 and there must be a path of length three
from v1 to v4.
5. Draw the directed graph that corresponds to the following adjacency
matrix:

Also, write down the adjacency list corresponding to the graph.

6. Write a C function to compute the indegree and outdegree of a vertex of a
directed graph when the graph is represented by an adjacency list.
7. Write a program in C to implement graph using adjacency matrix and
perform the following operations:

a. Depth First Search

b. Breadth First Search
2. Consider a graph representing an airlines network: each node represents an
airport, and each edge represents the existence of a direct flight between
two airports. Suppose that you are required to write a program that uses
this graph to find a flight route from airport p to airport q. Would you use
breadth first search or depth first search? Explain your answer.
3. Write the adjacency matrix of the graph shown in figure below. Does the
graph have a cycle? If it has, indicate all the cycles.

Fig. 8.47
4. Write a function that inserts an edge into an undirected graph represented
using an adjacency matrix.
5. Write a function that inserts an edge into a directed graph represented
using an adjacency matrix.
6. Write a function that deletes an edge from a directed graph represented
using an adjacency matrix.
7. Write a function that deletes an edge from an undirected graph represented
using an adjacency matrix.
8. Find the minimum spanning tree for the graph given in Figure 8.39.
9. Write the Warshalls algorithm.
10. Write the Dijkstras algorithm.

Chapter 9
Files
CHAPTER OUTLINE

9.1 Data and Information

9.2 File Concepts
9.3 File Organization
9.4 Files in C
9.5 Files and Streams
9.6 Working with Files Using I/O Stream
9.7 Sequential File Organization
9.8 Direct File Organization
9.9 Indexed Sequential Organization
9.10 Choice of File Organization
9.11 Graded Problems
9.1 DATA AND INFORMATION

The ancient scratches on the rocks, writings on Ashoka pillars, the

scriptures, and the caves of Ajanta and Elloraall indicate that the
human beings have a basic tendency to record information. The oldest
records were made on clay tablets as long ago as 3700 BC. Most of the
Indian scriptures were written on the thin sheets of bark of Bhojpatra, a
tree found in the upper Himalayas. In second century AD, paper was
developed in China. This event brought a revolution and the paper
became a writing medium in whole of the world.
In an enterprise, whether it is a small shop at the corner of the street, a
school or a large industry, it can be observed that each and every person
uses lot of paper and a large amount of data and information exchange
takes place. The workers deal with actual physical work and produce
data whereas the managers deal with the information written on papers.
Let us now delve upon the terms data and information, especially in the
context of files.

9.1.1 Data
The word data is a plural of datum, which means fact. Thus, data is a
collection of facts and figures. It can be represented by symbols. For
example, in a business house, data can be the name of an employee, his
salary, number of hours worked, etc. Similarly in an educational
institute, the data can be the marks of a student, roll numbers,
percentage, etc.
9.1.2 Information
It is the data arranged in an ordered and useful form. Thus, the data can
be processed so that it becomes meaningful and understandable
(see Figure 9.1).

Fig. 9.1 Processing of data into information

For example, in Figure 9.2(a), the collection of characters and digits is
meaningless by itself because it could refer to total marks and names of
students, the house numbers and names of house owners, etc. Once the
data is arranged in an order as shown in Figure 9.2(b), anybody can
come to know that the data refers to names of persons and their
corresponding ages. Therefore, this meaningful data can be called
information.

Fig. 9.2 Data and information

9.2 FILE CONCEPTS

Let us consider the information recorded in Figure 9.3. Each entry into
the table shown in this figure is a set of three data items: Roll, Name, and
Marks. The first entry states that the roll number and total marks of a
student named Ajay are 001 and 50, respectively. Similarly, the second
entry relates to a student named RAM whose roll number is 002 and
total marks are 67.

Each entry of the table is also known as a record of a student. The

record can be defined as a set of related data items of a particular entity
such as a person, machine, place, or operation. An individual data item
within a record is known as a field.
Thus, the record of a student is made up of following three fields:
1. Roll
2. Name
3. Marks

Let us assume that a particular class has 30 students. Now, the entire
information of the class consisting of 30 records is referred to as a file. A
file, therefore, consists of a number of records and each record consists
of a number of items known as fields. The creation and maintenance of
files is one of the most important activities in an organization.
Depending upon its purpose, a file can be called by a specific name or
type. The various types of files are tabulated in Table 9.1.

Fig. 9.3 File of records

Table 9.1 Types of files
Type

Purpose

Master file

Collection of records about an important activity. It may reflect the history

of events or information which can be used as a source of reference. The
master file is always kept up-to-date. These are fairly permanent files.

Stu
Per
Inv
Sto

Transaction
file

A temporary file with two purposes:

(1) collection of data about events as they occur and
(2) updation of master files.
These files are temporary in nature as these are deleted after master file is
updated.

Fee
Pur
Sal

Reference
file

This file contains information which is required for reference and changes
periodically, i.e., price list, students attendance, share rates, etc.

Stu
file
Sha

Report file

Data extracted from master files in order to prepare a report.

Stu
Sch
file

Sort file

A working file of records to be ordered in a particular sequence.

Me

Back up file

A copy of a master or a transaction file made to ensure that a duplicate is

available, if anything happens to the original.

Stu
bac

Archival
files

Copies made for long-term storage of data maintained for future reference.
These files are generally stored off-line, i.e., away from the computer
centre.

Sec
file

The arrangement of records in a file can be different for different

situations. The file of records shown in Figure 9.3 can be arranged
according to names instead of roll numbers, i.e., in alphabetical order.
The rearranged file is shown in Figure 9.4.
It may be noted here that in the file shown in Figure 9.3, the records
were arranged according to Roll field whereas in the file shown in Figure
9.4, the records are arranged according to Name field. However, a record
in both the files can be identified by the field called Roll. Basically, a field
which is used to identify a particular record within a file is called as
a key field. It is the key field according to which, generally, all records
within a file are arranged.

Fig. 9.4 File in alphabetical order

The records in a file can be arranged in the following three ways:
1. Ascending/Descending order: The records in the file can be arranged
according to ascending or descending order of a key field. For example, the
records in the file of Figure 9.3 are in ascending order of its Roll field.
2. Alphabetical order: If a field is of string or character type, then this
arrangement is used. For example, the records in the file of Figure 9.4 are in
alphabetical order.

3. Chronological order: In this type of order, the records are stored in the
order of their occurrence, i.e., arranged according to dates or events. If the
key field is a date, i.e., date of birth, date of joining, etc., then this type of
arrangement is used. For example, the records of a file shown inFigure
9.5 are arranged according to date of birth (DOB).

Fig. 9.5 File in chronological order

9.3 FILE ORGANIZATION

In a non-computerized company, the data and information are held on

files kept in a paper filing system. If the company had a computer, the
files would be held on computer storage. The former system is known as
manual filing system and the latter is known as electronic data
processing (EDP) filing system. The characteristics of a manual filing
system are as follows:
1. Records are stored as documents in files using papers.
2. Records are arranged using a system of indices.
3. The files are handled and maintained by people.

The characteristics of EDP filing system are as follows:

1.
2.
3.
4.

Used in computer environment.

Files are stored on computer storage.
Records are sorted and arranged according to a key field.
The files are handled and maintained by specific programs written in a high
level language such as PASCAL, C, software packages, etc.

The manual filing system is suitable for small organizations where fast
processing and storage of data is not required. In the EDP filing system,
large amount of data can be managed efficiently. The data storage and
retrieval becomes very fast. It may be noted that in the manual filing
system, the files are generally arranged in a meaningful sequence. The
records are located manually by the person in charge of the files.
However when the files are stored on the computer, the files have to be
kept in such a way that the records can be located, processed, and
selected easily by the computer program. The handling of files depends
on both input and output requirements, and the amount of data to be
stored. How best can the files be arranged for ease of access? This

necessity has led to the development of a number of file organization

techniques as listed below:
1. Sequential file organization
2. Direct access file organization
3. Indexed sequential ile organization
9.4 FILES IN C

All the programs that we have written so far have extensively used input
and output statements. The scanf statement is used to input data from
keyboard and printf statement to output or display data on visual display
unit (VDU) screen. Thus, whenever some data is required in a program,
it tries to read from the keyboard. The keyboard is an extremely slow
device. For small amount of data, this method of input works well. But
what happens when huge data is needed by a program? For example, a
program which generates the merit list of joint entrance examination
(JEE) may require data in the tune of 10,000 to 20,000 records. It is not
possible to sit down and type such a large amount of data in one go.
Another interesting situation is: what happens when the data or results
produced by one program are required subsequently by another
program? On the next day, the results may even be required by the
program that produced them. Therefore, we need to have a mechanism
such as files by virtue of which the program can read data from or write
data on magnetic storage medium.
When the data of a file is stored in the form of readable and printable
characters, then the file is known as a text file. On the other hand, if a file
contains non-readable characters in binary code then the file is called a
binary file. For instance, the program files created by an editor are stored
as text files whereas the executable files generated by a compiler are
stored as binary files. An introduction to files supported by C is given in
the following sections.
9.5 FILES AND STREAMS

In C, a stream is a data flow from a source to a sink. The sources and

sinks can be any of the input/output devices or files. For input and
output, there are two different streams called input stream and output
stream as shown in Figure 9.6.
We have already used the following formatted stream I/O functions in
our previous programs:
Stream

Description

scanf

standard input stream

printf

standard output stream

Fig. 9.6 Stream I/O

The standard source and sink are keyboard and monitor screen,
respectively. These unformatted streams are initialized whenever the
header file <stdio.h> is included in a program.
In fact, C supports a very large number of I/O functions capable of
performing a wide variety of tasks. The I/O system of C is categorized
into following three categories:
1. The stream I/O system
2. The console I/O system
3. The low-level I/O system

Some of the important stream I/O functions are listed in Table 9.2.
Table 9.2 Some important stream I/O functions in C
Function

Purpose

fclose

close a stream

feof

test for end of file

flush

flush a stream

fgctc

read a character from a stream

fgetchar

read a character from a stream

fgets

read a string from a stream

fopen

open a stream

fprintf

send formatted data to a stream

fputc

send a character to a stream

fputs

send a string to a stream

fread

read a block of data from a stream

fscanf

read formatted data from a stream

fseek

reposition a stream pointer

fwrite

write a block of data to a stream

fputs

send a string to a stream

getw

read an integer from a stream

putw

send an integer to a stream

Some of the important console I/O functions are listed in Table 9.3.
Table 9.3 Some important console I/O functions in C
Function

Purpose

cgets

read a string from the console

clrscr

clear text window

cprintf

write formatted data to the console

getch

read a character from the console

getche

read a character from the console and echo it

gotoxy

move the cursor to a specified location

putch

write a character to the console

Some of the low-level I/O functions are listed in Table 9.4.

Table 9.4 Some important low-level I/O functions in C
Function

Purpose

chmode

Change the mode a file

close

Close the file

eof

Test for end of file

filelength

Determine the length of a file in byte

lseek

Move or read data of a file pointer

open

Open a file

read

Read data from a file

setmode

Set mode of a file

write

Write data to a file

9.6 WORKING WITH FILES USING I/O STREAM

An I/O stream is a sequence of characters f lowing from a source to a

sink. When the data is read by a program from a device, the device
becomes the source which places the data on the stream (see Figure 9.6).
Obviously, the program becomes the sink that receives the data from the
stream. While writing the data, the program and the device reverse the
roles.
A stream is accessed by a stream pointer as shown in Figure 9.7. The
stream pointer points to the beginning of a stream.
A file is opened as a stream with the help of a pointer of type FILE.
Having opened the file, it is accessed through a pointer in a fashion very
similar to accessing an array of strings.

Fig. 9.7 Working of a stream

9.6.1 Opening of a File
A file can be opened in stream I/O by using function called fopen(). It
takes two arguments: file name and type of operation (see Table 9.5).
The fopen() function returns a pointer of type FILE.

Thus, a file (say Myfile) can be opened for reading by using the
following steps:
STEP 1: Declare a pointer (say ptr) of type FILE i.e., FILE*ptr;
STEP 2: Open the file using a stream I/O function called fopen().
ptr = fopen (Myfile, r);
The above statement requests the system to open a file called myfile
in read mode and assigns its pointer to ptr.
A file can be closed by a function called fclose() which takes one
argument of type FILE.
Thus, the file (Myfile) can be closed by the following statement:
fclose(ptr);
Table 9.5 File opening modes
Mod
e

Purpose

Open a file for reading. If file does not exist, NULL is returned.

Open a file for writing. It the file does not exist, a new file is created. If the file exists, the ne
overwrites the previous contents.

r+

Open the file for both reading and writing. If the file does not exist, NULL is returned.

w+

Open the file for both reading and writing. If the file exists, the previous contents are overw
one.

Open a file for appending. If the file exists, the new data is written at the end of the file else
created.

a+

Open the file for reading and appending. If the file does not exist, a new file is created.

9.6.2 unformatted File I/O Operations

An opened file can be read or written by the following set of unformatted
I/O functions:
fget(),get(),fgetchar()
getw(),fputc(),putc()
fputchar(),fputs(),putw()

A brief description of each function has already been provided in Table

9.2.
Let us consider a situation wherein we desire to create a file called
message.dat containing the following text:
When one apple fell, Newton was disturbed;
but when he found that all apples fell;
it was gravitation and he was satisfied.
The file message.dat would be opened with the help
of fopen() function and the functionfputs() would be used to write a
line in the file. The fputs() function requires two arguments as shown
below:
fputs(<string>,<filepointer>)
where <string> is string of characters to be written in the file; and
<filepointer> is the pointer to the file in which the text is to be
written.
The required program is given below:
#include<stdio.h>
main()
{
FILE*ptr;
ptr=fopen(message.dat,w);
if(!ptr)
{
printf(\nThefilecannotbeopened);
exit(1);
}
fputs(Whenanapplefell,Newtonwasdisturbed\n,
ptr);
fputs(butwhenhefoundthatallapplesfell,\n,
ptr);
fputs(itwasgravitationandhewassatisfied,
ptr);
fclose(ptr);
}
The above program is simple wherein the first statement of
function main() declares a pointer to aFILE and in the second
statement a file called message.dat is opened in write mode. The

next compound statement checks whether the operation was successful

or not. Rest of the statements write the text, line by line in the file. The
last statement closes the file.
A string from a file can be read by a function called fgets(). This
function requires three arguments as shown below:
fgets(<string>,n,<filepointer>)
where <string> is the character string, i.e., an array of char in which a
group of characters from the file would be read.
n is the number of characters to be read from the file. The function reads
a string of characters until the first new line (\n) character is
encountered or until the number of characters read is equal to n 2 1.
<filepointer> is the pointer to the file from which the text is to be
read.
The end of a file can be detected by a function called
feof(<file.pointer>). This function can be used in a situation where
a program needs to read whole of the file. For example, the following
program uses feof() and fgets() functions to read the
message.dat file till the end of the file is encountered. The strings
read from the file are displayed on the screen.
#include<stdio.h>
main()
{
chartext[80];
FILE*ptr;
ptr=fopen(message.dat,r);
if(!ptr)
{
printf(\nthefilecannotbeopened);
exit(1);
}
while(!feof(ptr))
{
fgets(text,80,ptr);
printf(\%s,text);
}
fclose(ptr);
}
Similarly, the functions fgetc() and fputc() functions can be used to
read from or write a character to a file. The function fputc() takes two

arguments, i.e., a character variable, and the file pointer whereas the
function fgetc() takes only one argument, i.e., the file pointer.
For example, ch=fgetc(ptr) reads a character in ch from a file
pointed by the pointer called ptr.fputc(ch,ptr) writes a character
stored in ch to a file pointed by the pointer called ptr.
Example 1: Write a program that copies the file
called message.dat to another file called new.dat.
Solution: The above program can be written by using the following
steps.
Steps
1. Open message.dat for reading.
2. Open new.dat for writing.
3. Read a character from message.dat. If the character is eof, then go to
step 5 else step 4.
4. Write the character in new.dat. Go to step 3.
5. Close message.dat and new.dat.

The required program is given below:

#include<stdio.h>
#include<conio.h>
main()
{
FILE*ptr1,*ptr2;
charch;
ptr1=fopen(message.dat,r);
if(!ptr1)
{
printf(\nThefile%scannotbeopened,
message.dat);
exit(1);
}
ptr2=fopen(new.dat,w);
if(!ptr2)
{
printf(\nThefile%scannotbeopened,
new.dat);
exit(1);
}
clrscr();

while(!feof(ptr1))
{
ch=fgetc(ptr1);
fputc(ch,ptr2);
}
fclose(ptr1);
fclose(ptr2);
}
The efficacy of the above program can be verified by opening the
new.dat either in notepad or in the Turbo-C editor.
It may be noted that till now we have used programs which are rigid in
nature, i.e., the programs work on particular files. For instance, the
above program works on message.dat and new.dat files and will
not work with other files. This situation can be easily handled by taking
the name of the file from the user at run time. Consider the program
segment given below:
.
.
.
charfilename[20];
FILE*ptr;
printf(\nEnterthenameofthefiletoopenfor
reading);
scanf(%s,filename);
ptr=fopen(filename,r);
.
.
.
The above program asks the user for the name of the file to be opened
giving freedom to the user to opt for any of the available files or to
provide altogether a new name for a file.
Example 2: Write a program that asks from the user for the file to be
opened and then displays the contents of the file.
Solution: The required program is given below:
#include<stdio.h>
#include<conio.h>
main()

{
charfilename[20],ch;
FILE*ptr;
printf(\nEnterthenameofthefiletobeopened
forreading:);
scanf(%s,filename);
ptr=fopen(filename,r);
if(!ptr)
{
printf(\nThefile%scannotbeopened,
filename);
exit(1);
}
clrscr();
while((ch=fgetc(ptr))!=EOF)
printf(%c,ch);
fclose(ptr);
}
Sampl
e

Input:

Enter the name of the file to be opened for

reading: new.dat

Output
:

When an apple fell, Newton was disturbed;

but when he found that all apples fell;
it was gravitation and he was satisfied.

It may be noted that in the above program, instead of feof(), the

keyword EOF has been used to detect the end of file. This is another
method of doing the same thing. In fact, it is especially useful where a
file has to be read character by character as was the case in the above
program.
Example 3: Write a program that reads a file of numbers and sorts the
numbers in ascending order. The sorted numbers are stored in a new
file.
Solution: Assume that the input file (say datafile.dat contains
the numbers which need to be sorted. The getw() function would be
used to read the numbers from the file into an array of integers

called List. The sorted list of numbers would be finally stored in an

output file (say sortedfile.dat).
The required program is given below:
/*Thisprogramreadsafileofnumbers,sortsthem
andwritesthesortedlistintoanewfile*/
voidsort(intList[100],intN);
#include<stdio.h>
main()
{
FILE*infile,*outfile;
intList[100];
inti,j,Num;
charinput_file[20],output_file[20];
printf(\nEnterthenameofInputfile:);
scanf(%s,input_file);
printf(\nEnterthenameofoutputfile:);
scanf(%s,output_file);
infile=fopen(input_file,r);
if(!infile)
{
printf(\nThefile%scannotbeopened,
input_file);
exit(1);
}
outfile=fopen(output_file,w);
if(!outfile)
{
printf(\nThefile%scannotbeopened,
output_file);
exit(1);
}
/*readtheinputfile*/
i=1;
while(!feof(infile))
{i++;
Num=getw(infile);
/*DonotincludeEOF*/
if(!feof(infile))List[i]=Num;
}

/*Sortthelist*/
sort(List,i);
/*writeintotheoutputfile*/
for(j=0;j<=i;j++)
putw(List[j],outfile);
fclose(infile);
fclose(outfile);
}
voidsort(intList[100],intN)
{
inti,j,small,pos,temp;
for(i=0;i<N1;i++)
{
small=List[i];
pos=i;
for(j=i+1;j<N;j++)
{
if(List[j]<small)
{
small=List[j];
pos=j;
}
}
temp=List[i];
List[i]=List[pos];
List[pos]=temp;
}
}
Note: The files created in the above program cannot be verified by
opening them in normal text editor like notepad. The reason is that the
functions getw() and putw() read/write the numbers in a file with a
format different from text format.
Therefore, a file consisting of numbers can be read by getw() function.
In fact, following program can be used to create the datafile.dat.
#include<stdio.h>
main()
{
FILE*ptr;
intval,i;
ptr=fopen(datafile.dat,w);

do
{
printf(\nEnteravalue);
scanf(%d,&val);
if(val!=9999)putw(val,ptr);
}
while(val!=9999);
fclose(ptr);
}
The numbers are written in the file as long as system does not encounter
9999, a number used to identify the end of the list.
Similarly, the following program can be used to see the contents of a file
(say sortedfile.dat)
#include<stdio.h>
main()
{
FILE*ptr;
intval,i;
ptr=fopen(sortedfile.dat,r);
printf(\n);
while(!feof(ptr))
{
val=getw(ptr);
if(!feof(ptr))printf(%d,val);/*Donotprint
EOF*/
}
fclose(ptr);
}
It may be further noted that both the programs take care that the EOF (a
number) does not get included into the list of numbers, being
manipulated by them.
9.6.3 Formatted File I/O Operations
C supports fscanf() and fprintf() functions to read or to write
from files. The general format of these functions is given below:
fscanf(<filepointer>,<formatstring>,<argument
list>);

fprintf(<filepointer>,<formatstring>,<argument
list>);
where <filepointer> is the pointer to the file in which I/O
operations are to be done.
<formatstring> is the list of format specifiers, i.e., %d, %f, %s, etc.
<argumentlist> is the list of arguments.
Example 4: Write a program that takes a list of numbers from the user
through standard input device, i.e., keyboard. The number 9999 marks
the end of the list. The list is then written on a user defined file by
using fprintf() function.
Solution: The required program is given below:
#includestdio.h
main()
{
FILE*ptr;
intval,i;
charoutfile[20];
printf(\nEnterthenameofthefile:);
scanf(%s,outfile);
ptr=fopen(outfile,w);
do
{
printf(\nEnteravalue);
scanf(%d,&val);
if(val!=9999)fprintf(ptr,%d,val);
}
while(val!=9999);
fclose(ptr);
}
Example 5: Use the file created in Example 4. Read the list of numbers
contained in the file usingfscanf() function. Print the largest number
in the list.
Solution: The numbers from the file would be read one by one and
compared for getting the largest of them. After EOF, the largest would be
displayed.
The required program is given below:

#includestdio.h
main()
{
FILE*ptr;
intval,i,large;
charinfile[20];
printf(\nEnterthenameofthefile:);
scanf(%s,infile);
ptr=fopen(infile,r);
printf(\nThelistis...);
large=9999;
while(1)
{
fscanf(ptr,%d,&val);
if(feof(ptr))break;
printf(%d,val);
if(large<val)large=val;
}
printf(\nThelargestis:%d,large);
fclose(ptr);
}
Sample output:
Enterthenameofthefile:Mydata.dat
Thelistis1234567832412113
Thelargestis:78.
9.6.4 Reading or Writing Blocks of Data in Files
A block of data such as an array or a structure can be written in a file by
using fwrite() function. Similarly, the block of data can be read from
the file using fread() function.
The general formats of fread() and fwrite() functions are given
below:
fread(<addressofobject>,<sizeofobject>,<number
ofitems>,<filepointer>);
fwrite(<addressofobject>,<sizeofobject>,<number
ofitems>,<filepointer>);
where <addressofobject> is the address or pointer to the object
from which the data is to be read/ written on the file and vice versa.

<sizeofobject> is computed by the function sizeof() and

included in the argument list
<numberofitems> is the number of data items to be read or written.
Generally, it is set as 1.
<filepointer> is the pointer to the file from which data is to be read
or written.
Let us now try to write the following structure into a file
(say test.dat):
structstud{
charname[20];
introll;
};
The following statements can be used to do the required task:
FILE*ptr;
structstud{
charname[20];
introll;
};
structstudob={Ram,101};
ptr=fopen(test.dat,w);
fwrite(&ob,sizeof(structstud),1,ptr);
fclose(ptr);
A similar code can be written to read the structure from a file.
A complete program that writes a structure into a file and then reads and
displays it, is given below:
#include<stdio.h>
main()
{FILE*ptr;
structstud{
charname[20];
introll;
};
structstudob1;
structstudob={Ram,101};
ptr=fopen(test.dat,w);
fwrite(&ob,sizeof(structstud),1,ptr);
fclose(ptr);

ptr=fopen(test.dat,r);
fread(&ob1,sizeof(structstud),1,ptr);
fclose(ptr);
printf(\n%s,ob1.name);
printf(\n%d,ob1.roll);
}
Example 6: Write a program that creates a file
called Marks.dat and writes the records of all the students studying
in a class.
Solution: The required program is given below:
#include<stdio.h>
structstudent{
charname[20];
introll;
intmarks;
};
main()
{FILE*ptr;
structstudentstudob;
intt=sizeof(structstudent);
ptr=fopen(marks.dat,w);
printf(\nEnterthestudentdataterminatedbyroll
=9999);
while(1)
{
printf(\nName:);scanf(%s,studob.name);
printf(\nRoll:);scanf(%d,&studob.roll);
if(studob.roll==9999)break;
printf(\nmarks:);scanf(%d,&studob.marks);
fwrite(&studob,t,1,ptr);
}
fclose(ptr);
}
Example 7: A file called Marks.dat contains the records of students
having the following structure:
structstudent
{
charname[20];

introll;
intmarks;
};
Write a program that reads marks and creates the following two files:
1. Pass.dat should contain the roll numbers of students scoring more than
or equal to passmarks stored in a variable called pass_marks.
2. FAIL.dat should contain the roll numbers of students scoring below the
pass_marks.

Solution: The required program is given below:

#include<stdio.h>
structstudent{
charname[20];
introll;
intmarks;
};
main()
{FILE*ptr,*p,*f;
structstudentstudob;
intpass_marks;
intt=sizeof(structstudent);
ptr=fopen(marks.dat,r);
p=fopen(pass.dat,w);
f=fopen(fail.dat,w);
printf(\nEnterthepassingparks:);
scanf(%d,&pass_marks);
while(!feof(ptr))
{
fread(&studob,t,1,ptr);
if(feof(ptr))break;
if(studob.marks>=pass_marks)
fprintf(p,%d,studob.roll);
else
fprintf(f,%d,studob.roll);
}
fclose(ptr);
fclose(p);
fclose(f);
}

Example 8: Write a program that opens the pass.dat and

fail.dat files created in the program in Example 7 and displays two
separate lists of roll numbers who have passed and failed in the
examination.
Solution: The required program is given below:
#include<stdio.h>
main()
{introll;
FILE*p,*f;
p=fopen(pass.dat,r);
if(!p)
{
printf(\nThefile%scannotbeopened,
pass.dat);
exit(1);
}
f=fopen(fail.dat,r);
if(!f)
{
printf(\nThefile%scannotbeopened,
fail.dat);
exit(1);
}
printf(\nThelistofpassstudents:);
while(!feof(p))
{
fscanf(p,%d,&roll);
if(!feof(p))printf(%d,roll);
roll=getch();
}
printf(\nThelistoffailstudents:);
while(!feof(f))
{
fscanf(f,%d,&roll);
if(!feof(f))printf(%d,roll);
}
fclose(p);
fclose(f);
}
}

9.7 SEQUENTIAL FILE ORGANIZATION

This file organization is the simplest way to store and retrieve records of
a file. In this file, the records are stored in a meaningful order according
to a key field. The first record in the order is placed at the beginning of
the file. The second record is stored right after the first, the third after
the second, and so on. However, the records may be ordered in
ascending or descending order by the key field which can be numeric
(such as student roll number) or alphabetic (such as student name). It
may be noted that the order never changes afterwards. The arrangement
of records in a sequential file is shown in Figure 9.8.
Since all records in the file are stored sequentially by position, the
records can be identified by the key field only.

Fig. 9.8 Arrangement of records in a file

9.7.1 Creating a Sequential File
The sequential file can be created on a storage device (magnetic tape or
disk) by the following three steps:
1. A name for the file is chosen. The system is requested to open a file on the
storage device with the chosen file name.
2. The records for the file are input sequentially one by one into the computer
which in turn stores them in the order of their arrival on to the file.
3. After all the records are transferred into the file, the system is requested to
close the file.

A flow chart for this process is given in Figure 9.9. The above mentioned
steps are carried out with the help of a program written in a high level
language. All the high level languages, such as BASIC, FORTRAN,
PASCAL and C, support sequential files.
9.7.2 Reading and Searching a Sequential File
Suppose student data of an engineering college is maintained in a file
(say stud.dat)
Let us assume that on a particular day, the principal of the school
requires some information from the file called stud.dat. Now, the
obvious step is to read the file to get the required information. The read
operation for a sequential file is explained below.

To read a sequential file, the system always starts at the beginning of the
file and the records are read one by one till the required record is
reached or end of the file is encountered. The end of the file (EOF)
indicates that all the records in the file have been read. For instance, if
the desired record happens to be 50th one in a file, the system starts at
the first record and reads ahead one record at a time until the 50th is
reached.
The reading of a file involves the following operations:
1. Open the file.
2. Read a record.
3. Check if the record is the desired one. If yes, then display information and
perform step 6.
4. Check for the end of the file (EOF). If yes, then perform step 6.
5. Go to step 2.
6. Close the file.

A flow chart for the reading of file is given in Figure 9.10.

The step 3, i.e., check if the record is the desired one is carried out by
matching the key field of the record read from the file with the key field
of the desired record. For example, if the principal desires the
information about a student whose roll number is 125 then
the ROLL field of each record read from the file will be compared with
the value 125. The process will be repeated till a record with ROLL equal
to 125 is encountered.

Fig. 9.9 Creation of a sequential file

9.7.3 Appending a Sequential File
The append operation is done with a view of adding more records to an
existing file. In this operation, the file is opened and is read record by
record, till EOF is encountered. The new records which are to be added to
the file are then written into the file. Once all the records have been
added, the file is closed. Thus, the following operations are done for the
purpose of appending a file:
1.
2.
3.
4.
5.

Open the file for append operation.

Read the file till EOF is encountered.
Read the record to be added.
Write the record on the file.
Close the file.

Fig. 9.10 Reading sequential file

Fig. 9.11 Appending a sequential file

A flow chart for appending of a file is given in Figure 9.11.
Example 9: Write a program that creates a sequential file of records of
students who are players of different games. The structure of student
record is given below:
structstudent
{
charname[20]
introll;
charclass[5];
chargame[15];
};

Solution: The required program that creates the file (say

Play_Master.dat) is given below:
/*Thisprogramcreatesasequentialfile*/
#include<stdio.h>
structstudent{
charname[20];
introll;
charclass[5];
chargame[15];
};
main()
{
FILE*ptr;
charmyfile[15];
intt=sizeof(structstudent);
structstudentstudrec;
printf(\nEnterthenameofthefile:);
scanf(%s,myfile);
ptr=fopen(myfile,w);
printf(\nEntertherecordsofthestudentsoneby
one);
studrec.roll=0;
while(studrec.roll!=9999)
{
printf(\nName:);fflush(stdin);
gets(studrec.name);
printf(\nRoll:);scanf(%d,&studrec.roll);
printf(\nClass:);fflush(stdin);
gets(studrec.class);
printf(\nGame:);fflush(stdin);
gets(studrec.game);
if(studrec.roll!=9999)
fwrite(&studrec,t,1,ptr);
}
fclose(ptr);
}
Example 10: Write a program that uses Play_Master.dat created
in Example 9. For a given game, it displays the data of all the students
who play that game.

Solution: The required program is as follows:

/*Thisprogramsearchesasequentialfile*/
#include<stdio.h>
structstudent{
charname[20];
introll;
charclass[5];
chargame[15];
};
main()
{
FILE*ptr;
chargame[15],myfile[15];
intt=sizeof(structstudent);
structstudentstudrec;
printf(\nEnterthenameofthefiletobeopened);
scanf(%s,myfile);
ptr=fopen(myfile,r);
printf(\nEnterthenameofthegame:);
fflush(stdin);
gets(game);
fread(&studrec,t,1,ptr);
while(!feof(ptr))
{
if(!strcmp(game,studrec.game))
{printf(\nName=%s,studrec.name);
printf(\nRoll=%d,studrec.roll);
printf(\nclass=%s,studrec.class);
printf(\nGame=%s,studrec.game);
}
fread(&studrec,t,1,ptr);
}
fclose(ptr);
}
Example 11: Write a program that opens the Play_Master.dat file
and appends records in the file till the roll = -29999 is encountered.
Solution: The required program is given below:

/*Thisprogramappendsrecordsinasequentialfile
*/
#include<stdio.h>
structstudent{
charname[20];
introll;
charclass[5];
chargame[15];
};
main()
{
FILE*ptr;
charmyfile[15];
intt=sizeof(structstudent);
structstudentstudrec;
printf(\nEnterthenameofthefile:);
scanf(%s,myfile);
ptr=fopen(myfile,r+);
printf(\nEntertherecordstobeappended);
while(!feof(ptr))
fread(&studrec,t,1,ptr);
studrec.roll=0;
while(studrec.roll!=9999)
{
printf(\nName:);fflush(stdin);
gets(studrec.name);
printf(\nRoll:);scanf(%d,&studrec.roll);
printf(\nClass:);fflush(stdin);
gets(studrec.class);
printf(\nGame:);fflush(stdin);
gets(studrec.game);
if(studrec.roll!=9999)
fwrite(&studrec,t,1,ptr);
}
fclose(ptr);
}
It may be noted that in the above program, the file was opened in r+
mode. In fact, the same job can be done by opening the file in a, i.e.,
append mode. The program that uses this mode is given below:

/*Thisprogramappendsrecordsinasequentialfile
*/
#include<stdio.h>
structstudent{
charname[20];
introll;
charclass[5];
chargame[15];
};
main()
{
FILE*ptr;
charmyfile[15];
intt=sizeof(structstudent);
structstudentstudrec;
printf(\nEnterthenameofthefile:);
scanf(%s,myfile);
ptr=fopen(myfile,a);
printf(\nEntertherecordstobeappended);
studrec.roll=0;
while(studrec.roll!=9999)
{
printf(\nName:);fflush(stdin);
gets(studrec.name);
printf(\nRoll:);scanf(%d,&studrec.roll);
printf(\nClass:);fflush(stdin);
gets(studrec.class);
printf(\nGame:);fflush(stdin);
gets(studrec.game);
if(studrec.roll!=9999)
fwrite(&studrec,t,1,ptr);
}
fclose(ptr);
}
Example 12: Modify the program written in Example 11 so that it
displays the contents of a file before and after the records are appended
to the file.
Solution: A function called show() would be used to display the
contents of the file. The required program is given below:

/*Thisprogramdisplaysthecontentsbeforeand
afteritappendsrecordsinasequentialfile*/
#include<stdio.h>
voidshow(FILE*p,charx[15]);
structstudent{
charname[20];
introll;
charclass[5];
chargame[15];
};
main()
{
FILE*ptr;
charmyfile[15];
intt=sizeof(structstudent);
structstudentstudrec;
printf(\nEnterthenameofthefile:);
scanf(%s,myfile);
printf(\nTheRecordsbeforeappend);
show(ptr,myfile);
ptr=fopen(myfile,a);
printf(\nEntertherecordstobeappended);
studrec.roll=0;
while(studrec.roll!=9999)
{
printf(\nName:);fflush(stdin);
gets(studrec.name);
printf(\nRoll:);scanf(%d,&studrec.roll);
printf(\nClass:);fflush(stdin);
gets(studrec.class);
printf(\nGame:);fflush(stdin);
gets(studrec.game);
if(studrec.roll!=9999)
fwrite(&studrec,t,1,ptr);
}
fclose(ptr);
printf(\nTheRecordsafterappend);
show(ptr,myfile);
}
voidshow(FILE*p,charmyfile[15])

{
structstudentstudrec;
p=fopen(myfile,r);
fread(&studrec,sizeof(studrec),1,p);
while(!feof(p))
{
printf(\nName:);fflush(stdout);
puts(studrec.name);
printf(\nRoll:);printf(%d,studrec.roll);
printf(\nClass:);fflush(stdout);
puts(studrec.class);
printf(\nGame:);fflush(stdout);
puts(studrec.game);
fread(&studrec,sizeof(studrec),1,p);
}
fclose(p);
}
9.7.4 updating a Sequential File
The updation of a file means that the file is made up-to-date with the
latest changes relating to it. The original file is known as a master file or
old master file and the temporary file which contains the changes or
transactions is known as transaction file. Examples of transactions are
making purchases, payment for purchases, total sales in the day, etc.
In order to update a sequential file, the transactions are sorted according
to the same key used for the old master file, i.e., both the files are sorted
in the same order on the same key. Thus, the master file and the
transaction file become co-sequential. A new file called new master file
is opened for writing. The old master file and the transaction file are
merged according to the following logic:
The key of the record from the transaction file is compared with key of
the record from the old master file. If identical, the transaction record
is written on the new master file. Otherwise, the master record is
written on the new master file.

Fig. 9.12 Transaction processing

Example 13: A file called old master of a bank contains the records of
saving fund A/C of its customers in ascending order of account numbers.
The format of the record is given below:
Name
Acnt_No
Balance_Amt
A file called transaction file contains the details about transactions done
in A/C on a particular day. It also contains the records in the ascending
order of A/C numbers. Each record of that file has the following format:
Acnt_No
Trans_type
Trans_amt
The field Trans-amt contains the amount withdrawn or deposited to the
account depending upon the value of Trans-type being 0 or 1,
respectively. Write a program that reads the two files and does the
transaction processing giving a third file called new master.
Solution: In this program, the following menu would be displayed so
that the possible actions can be conveniently carried out by the user:
Menu
Createoldmasterfile

Createtransactionfile

Generatenewmasterfile

Showcontentsofafile

Quit

The required program is given below:

/*ThisprogrammaintainsthesavingA/Cofabank*/
#include<stdio.h>
#include<conio.h>
structAcnt_Rec{
charname[20];
intAcnt_No;
floatBalance_Amt;
};
structTrans_Rec{
intAcnt_No;
intTrans_type;
floatTrans_Amt;
};
voidcreat_T(charTrans[15]);
voidcreat_O(charO_master[15]);
voidcreat_N(charO_master[15],charTrans[15],char
N_master[15]);
voidshowfile(charfname[15],intf_type);
main()
{
structAcnt_RecAcntob;
structTrans_RecTransob;
charO_master[15],N_master[15],Trans[15],
fname[15];
intchoice,f_type;
/*ShowMenu*/
do
{clrscr();
printf(\nMenu);
printf(\n);
printf(\nCreateOldmasterfile0);
printf(\nCreateTransactionfile1);
printf(\nCreateNewmasterfile2);
printf(\nShowContentsofafile3);

printf(\nQuit4);
printf(\nEnteryourchoice:);scanf(%d,
&choice);
switch(choice)
{
case0:clrscr();
printf(\nEnterthenameofthefile);
fflush(stdin);gets(O_master);creat_O(O_master);
break;
case1:clrscr();
printf(\nEnterthenameofthefile);
fflush(stdin);gets(Trans);creat_T(Trans);
break;
case2:clrscr();
printf(\nEnterthenameofOldmasterfile);
fflush(stdin);gets(O_master);
printf(\nEnterthenameofNewmasterfile);
fflush(stdin);gets(N_master);
printf(\nEnterthenameofTransactionfile);
fflush(stdin);gets(Trans);
creat_N(O_master,Trans,N_master);
break;
case3:printf(\nEnterthenameofthefiletobe
displayed);
fflush(stdin);gets(fname);
printf(\nEnter0/1forMaster/Transaction);
scanf(%d,&f_type);showfile(fname,f_type);
break;
}
}
while(choice!=4);
}
voidcreat_O(charO_master[15])
{
FILE*ptr;
structAcnt_RecAob;
intt=sizeof(Aob);
ptr=fopen(O_master,w);
printf(\nEntertherecordsonebyoneterminatedby
A/c=9999);
Aob.Acnt_No=0;

while(Aob.Acnt_No!=9999)
{
printf(\nName:);fflush(stdin);gets(Aob.name);
printf(\nA/CNo:);scanf(\n%d,&Aob.Acnt_No);
printf(\nBalanceAmount:);scanf(\n%f,
&Aob.Balance_Amt);
if(Aob.Acnt_No!=9999)fwrite(&Aob,t,1,ptr);
}
fclose(ptr);
}
voidcreat_T(charTrans[15])
{FILE*ptr;
structTrans_RecAob;
intt=sizeof(Aob);
ptr=fopen(Trans,w);
printf(\nEntertherecordsonebyoneterminated
byA/c=9999);
Aob.Acnt_No=0;
while(Aob.Acnt_No!=9999)
{
printf(\nA/CNo:);scanf(\n%d,&Aob.Acnt_No);
printf(\nTransactionType(0/1):);scanf(\n
%d,&Aob.Trans_type);
printf(\nTransactionAmount:);scanf(\n%f,
&Aob.Trans_Amt);
if(Aob.Acnt_No!=9999)fwrite(&Aob,t,1,ptr);
}
fclose(ptr);
}
voidcreat_N(charO_master[15],charTran[15],char
N_master[15])
{
FILE*Old,*New,*Trans;
structAcnt_RecOob;
structTrans_RecTob;
intt1,t2;
t1=sizeof(structAcnt_Rec);
t2=sizeof(structTrans_Rec);
/*openfiles*/
Old=fopen(O_master,r);

Trans=fopen(Tran,r);
New=fopen(N_master,w);
/*performtransactionprocessing*/
fread(&Oob,t1,1,Old);
fread(&Tob,t2,1,Trans);
while(!feof(Old)&&!feof(Trans))
{
if(Oob.Acnt_No<Tob.Acnt_No)
{fwrite(&Oob,t1,1,New);
fread(&Oob,t1,1,Old);
}
else
{
while((Oob.Acnt_No==Tob.Acnt_No)&&!
feof(Trans))
{
if(Tob.Trans_type==0)
Oob.Balance_Amt=Oob.Balance_Amt
Tob.Trans_Amt;
else
Oob.Balance_Amt=Oob.Balance_Amt+Tob.Trans_Amt;
fread(&Tob,t2,1,Trans);
}
fwrite(&Oob,t1,1,New);
fread(&Oob,t1,1,Old);
}
}
/*Copyrestofthefile*/
while(!feof(Old))
{
fwrite(&Oob,t1,1,New);
fread(&Oob,t1,1,Old);
}
/*Closefiles*/
fclose(Old);
fclose(New);
fclose(Trans);
}
voidshowfile(charfname[15],intf_type)
{FILE*ptr;
charch;

structAcnt_RecAob;
structTrans_RecTob;
intt1=sizeof(Aob);
intt2=sizeof(Tob);
if(f_type==0)
{
ptr=fopen(fname,r);
fread(&Aob,t1,1,ptr);
while(!feof(ptr))
{
printf(\nName:);fflush(stdout);
puts(Aob.name);
printf(\nA/CNo:);printf(\n%d,
Aob.Acnt_No);
printf(\nBalanceAmount:);printf(\n%f,
Aob.Balance_Amt);
fread(&Aob,t1,1,ptr);
ch=getch();
}
}
else
{ptr=fopen(fname,r);
fread(&Tob,t2,1,ptr);
while(!feof(ptr))
{
printf(\nA/CNo:);printf(\n%d,
Tob.Acnt_No);
printf(\nTransactionType:);printf(\n%d,
Tob.Trans_type);
printf(\nTransactionAmount:);printf(\n%f,
Tob.Trans_Amt);
fread(&Tob,t2,1,ptr);
ch=getch();
}
}
fclose(ptr);
}

9.8 DIRECT FILE ORGANIZATION

The direct files allow records to be read or written on to a file without

proceeding sequentially from the beginning. In this organization, the
records are not stored in sequence. To read or write, an address is first
calculated by applying a mathematical function to the key field of the
record. The record is then stored at the generated address. The
mathematical function is called a Hash function.
However, a poor form of direct access can be obtained in C files by using
the fseek() function. With the help of this function, the stream file
pointer can be positioned at a particular record. It takes three arguments
as shown below:
fseek(<filepointer>,<offset>,<startposition>)
wher
e

<filepointer> is the pointer to the file.

<startposition> is the start position of the file. It can
have either of the following values:

Value

Meaning

The starting position is the beginning of the file.

Use the current file pointer as the starting position.

The starting position is the end of the file.

<Offset
>

is the offset, i.e., how much the pointer should advance

from the starting position.

From Figure 9.13, it may be noted that file pointer (ptr) is positioned
at the beginning of the file because the start position is equal to 0 (zero).
The offset (i.e., 4) would move the file pointer to the fourth record.

Fig. 9.13 The working of fseek(ptr,4,0) function.

Similarly, the fseek(ptr,9,2) suggests that from the end of the file
move back by 9 records.

Thus, it is a relative file organization wherein the records are

accessed directly in relation with the starting point, specified within the
file. The offset is usually computed by the following formula:
offset=recordnumbersizeofarecord
Example 14: Write a program that creates a file of room data of a hotel
having following structure. The rooms are numbered starting from 1.
structroom
{
introom_no;
charroomtype;
charstatus;
};
where room type may take value S/D, i.e., single bed or double bed.
Similarly, status may take value A/N, i.e., available or not available.
Use fseek() function to display the information about a particular
room of the hotel.
Solution: The required program is given below:
/*Thisprogramusesfseek()functiontosearch
recordsinafile*/
#include<stdio.h>
structroom{
introom_no;
charroom_type;
charroom_status;
};
main()
{FILE*ptr;
charfilename[20];
intt=sizeof(structroom);
structroomrob;
intno_of_rooms,i,choice;
printf(\nEnterthenameofthefile);
fflush(stdin);gets(filename);
printf(\nEnterthenumberofrooms);
scanf(%d,&no_of_rooms);
/*createthefile*/
ptr=fopen(filename,w);

printf(\nEntertheroomdataonebyone);
for(i=1;i<=no_of_rooms;i++)
{rob.room_no=i;
printf(\nRoomNo.%d:,i);
printf(\nRoom_Type(S/D):);
fflush(stdin);scanf(%c,&rob.room_type);
rob.room_type=toupper(rob.room_type);
printf(\nRoom_Status(A/N):);
fflush(stdin);scanf(%c,&rob.room_status);
rob.room_status=toupper(rob.room_status);
fwrite(&rob,t,1,ptr);
}
fclose(ptr);
ptr=fopen(filename,r);
/*searcharecord*/
do
{
printf(\nMenu);
printf(\n);
printf(\nDisplayinfo1);
printf(\nQuit2);
printf(\nEnteryourchoice);
scanf(%d,&choice);
if(choice==1)
{
printf(\nEntertheroomno.);
scanf(%d,&i);
fseek(ptr,(i1)*t,0);
fread(&rob,t,1,ptr);
clrscr();
printf(\nRoomNo.=%d,rob.room_no);
printf(\nRoomType=%c,rob.room_type);
printf(\nRoomStatus=%c,rob.room_status);
}
else
break;
}
while(1);
fclose(ptr);
}

The characteristics of a relative file are given below :

1. A relative file has fixed length records.
2. Al records in the file are written and accessed with the help of a record
number.
3. Unless the record number for a record within the file is known, one cannot
gain access to the information that the record contains.
4. The file is open for both writing and reading operations at once.

The advantages of direct files are given below :

1. Any record can be directly accessed.
2. File processing and updation activities can be performed without the
creation of new master file.
3. Speed of record processing in case of large files is very fast as compared to
sequential files.
4. The files can be kept up-to-date because on-line updation is possible.
5. Concurrent processing of several files is possible. For example, when a sale
is recorded in a company, the account receivable can be posted on one file
and simultaneously the inventory withdrawal can also be written on another
file.
9.9 INDEXED SEQUENTIAL ORGANIZATION

A very large number of applications require a mixture of sequential and

direct processing. For example, in an inventory control system, it is
desired that at regular intervals the transactions be processed
sequentially in a file and if needed, make direct accesses inquiries into
the same file. File organization using the Indexed Sequential Access
Method (ISAM) provide the essential benefits of both the sequential and
random methods.
The indexed organization is similar to a telephone directory wherein one
can retrieve a telephone number by using an index provided at the
beginning. In this organization, the records of the file are stored
sequentially in blocks on the disk. Each block can store a specified
amount or set of of records. The records in this file can be accessed
through an index. An index is a separate file from the original file. Each
entry into the index contains only two items of data: the value of the
keyfield of last record in the block and its corresponding starting block
address (see Figure 9.14 ).

Fig. 9.14 Indexed sequential organization

9.9.1 Searching a Record
To search a specified record in an indexed sequential file, the index is
first searched to find the block in which the key of the required record is
present. When index entry is found, the address of the corresponding
block containing the record on the disk is noted and then whole of the
block is brought in the main memory. Once the block is brought inside
the main memory, the record is searched sequentially in the block.
It may be noted from Figure 9.14 that each entry in the index contains
the highest key value for any record in the block and the starting address
of the block. Now if we want to search a record with key value 376, then
the index is searched first. The first entry (110) of the index is compared
with the search key (376). Since 110 is the highest key value in its
corresponding block and it is less than the search key (376), the
comparison with next entry is done. The comparison continues till it
reaches to third entry and it is found that the key of the entry (417) is
larger than the search key (376). Thus it is established that the required
record should be on its corresponding block. The blocks starting address
is 4127. Now the access is made to disk address 4127 and the block is
brought into the main memory. The search key is compared with key of
the first record of the block (370). Then it moves on to the second record
and the process is repeated till it reaches to the third record (key = 376).
The keys match establishing the fact that the required record has been
found. Now, it can perform the desired operation on the record such as
display the contents, print certain values, etc.

9.9.2 Addition/Deletion of a Record

When records are added to an already existing indexed sequential file,
the system inserts them at proper places so as to retain the sequential
nature of each block. For example, assume that records with keys 217
and 371 are to be added into the file shown in Figure 9.14. The record
with key 217 is searched into the file and its exact location where the
record is to be inserted i.e. in the block starting with address 4130 and
after the record with key value 216. Figure 9.15 shows that the record
with key 217 has been inserted after record with key 216 and the rest of
the records have been moved down to create space for the incoming
record.

Fig. 9.15 Operations on indexed sequential file

The addition of record (371) will push the record (417) out of the Block 3.
The reason being that the block is already full, and to accommodate the
incoming record (371) the record from the extreme end has to leave the
block. The overflowed record is then stored in a separate block known as
overflow block. The deletion of a record from the file is logical deletion,
i.e., the record is not physically or actually deleted from the file but it is
marked with special characters to indicate that the record is no longer
accessible. For example, if the records with key value 002 and 219 are to
be deleted then these records are searched in the file. If they are found
then the records are marked with special characters (< > in our case)
(see Figure 9.15 ). The presence of < > in a record indicates that the
record has been deleted.
The modifications in a record involves the following steps
1. Search the record
2. Read the record

3. Change the record

4. Rewrite the changed record in its original location.

The ISAM files are very flexible in the sense that they can be processed
either sequentially or randomly, depending on the requirements of the
user.
9.9.3 Storage Devices for Indexed Sequential Files
The magnetic disks are the most suitable storage media for Indexed
sequential files. The ISAM files are mapped on to the cylinders of the
disk. We know that the index file consists of three parts: index, main file
(data part), and overflow area. A cylinder on the disk is accordingly
divided into three parts: index area, prime area, and overflow area as
shown in Figure 9.16

Fig. 9.16 ISAM on a cylinder of a magnetic disk

1. Index area: The first track in each cylinder is reserved for an index. This
index describes the storage of records on the various tracks of the cylinder.
2. Prime area (Data tracks): The various records of the file are written on
this area on a cylinder always starting from the 2nd track. The records are
recorded sequentially along the various tracks. The highest key on the track
and the track address are entered into the index.
3. Overflow area: The area is created to accommodate the overflowed
records in the file. The overflow area is generally the last track on the same
cylinder. This area is unoccupied at the time of creation of file.

9.9.4 Multilevel Indexed Files

If the file is so large that it cannot be accommodated on one cylinder
then the records of the file are stored on many cylinders. To handle this
situation, a separate index known as cylinder index is maintained to
indicate how records are distributed over a number of cylinders.
Therefore, we have a hierarchy of indices in the form of multilevel index
as shown in Figure 9.17. The cylinder index has entries consisting of two
parts: the highest key value on the cylinder and the cylinder number.
To search a record in a multi level index file, the cylinder index is first
searched to find the cylinder on which the record is stored. Once the
cylinder is known then the track index is searched to find the block on
which the record is residing. Rest of the process is same as in the case of
single level indexed file.
The advantages of indexed files are as follows:
1. Provides good flexibility for users who need both direct accesses and
sequential progressing with the same file.
2. It makes rapid access possible over the indexes.
3. Interactive processing is possible.

The disadvantages of indexed files are as follows:

1. ISAM files generate excessive overflows, and vacant storage locations are
created.
2. Extra storage space is required for the index especially when multilevel
indexes are maintained.
3. The overflow areas sometimes overflow to other cylinders and thus cause
much read/write head movement making the access very slow. Therefore,
the file has to be reorganized and indexes rearranged accordingly. This is a
time consuming process.
9.10 CHOICE OF FILE ORGANIZATION

The choice of a file organization depends on the nature of a given

application. How a data file is used, determines which particular file
organization method is appropriate for it. The following factors play an
important part in the selection of a file organization.
1. Activity ratio: A record is said to be active if it requires processing in
a file. The portion containing records in a file that requires processing is
called as activity of a file. The ratio of number of active records and the
total number of records in the file is known as activity ratio

Fig. 9.17 Multilevel indexed file

A sequential file can be used if the activity ratio is high, i.e., more than
6070 per cent. On the other hand, if the activity ratio is very low, then
the direct file can be chosen. The indexed sequential files are less
efficient than the direct files for low activity. A comparison between the
various files organizations is shown in Figure 9.18.

Fig. 9.18 Impact of file activity ration on a file type

2. Response time: The response time can be defined as the amount of
time elapsed between a demand for processing and its completion. The
response time of a sequential file is always very long because it involves
both the waiting time i.e. until a batch of transaction is run and the
processing time. The direct and indexed sequential files have a faster
response time.
The direct files allow quick retrieval of records. Generally faster response
time is comparatively more expansive. Since the direct and indexed files
are more up to date, the added cost per transaction is justified.
3. Volatility: It is the measure of percentage of records added to or
deleted from a file during its use over a period of time. Since each file
organization requires different methods to add or delete records, the
volatility becomes a very important factor as far as file processing and its
organization is concerned.
In a direct file, the process for adding and deleting records is simple and
time saving compared to a sequential file. The indexed sequential files
are also very volatile. The direct or indexed files have a problem of
reorganization and overflow overheads.
4. Size: The amount of processing required for a file is directly
proportional to its size. The size of a file is determined in terms of total
number of records in the file. The size of a record is measured in terms of
total number of characters in the record.
Small files do not require any painstaking design. On the other hand
large ones require careful design. A small sequential file may be stored
on a direct access device to avoid waiting time, i.e., mounting and
dismounting of tapes is avoided. Large sequential files are to be stored
on a tape.* As far as direct files are concerned, the size has no relevance
5. Integration: Generally, a business organization maintains
integrated file system that allows online updation of interdependent
files. Random and indexed organization allow all the files affected by a

particular transaction to be updated online. Sequential files are not

suitable for such system of files.
6. Security: A back up is taken during every run of a sequential file.
Therefore a sequential file stored is more secured. In fact, sequential files
processing provide automatic data backup. The data stored in direct and
indexed files are comparatively not secured. Special techniques such as
frequent mirroring of disks, are used to insure file integrity. On the other
hand the magnetic tapes are more prone to physical damage.
9.11 GRADED PROBLEMS

Problem 1: A macro processor is a text substitution program where

macro consists of a macro name and its definition as shown below:
Macroname

Macrodefinition(replacementtext)

YMCA

YMCAinstituteofEngineering

The macro processor replaces all occurrences of macro name in a

document by its corresponding definition. Write a program that takes a
text in which the macros are defined at the beginning by the statements
such as given below. The definitions end by the keyword MEND.
#define

CE

ComputerEngineering

#define

PC

Personalcomputer

MEND
<textofthedocument>
The program implements the macro processor, i.e., reads the macros and
their definitions and does the text substitution for each macro call.
Solution: The following data structures would be employed to do the
required task:
1. A structure to store a macro name and its definition.
structmacro
{
charName[10];

charMdef[50];
}
2. An array of structures of type macro to store all macro definitions.
structmacroMlist[15]
The program would require the following functions:

Function

Description

1.gettaken()

toreadawordfromtheinput
textdocument.

2.build
Mtable()

tobuildatableofmacronames
anddefinitions
usingthearrayofstructure,
i.e.,Mlist[].

3.writeout()

writetheoutputtotheoutput
textfile.

A complete program that does the required task is given below:

/*Thisprogramimplementsamacroprocessor,atext
substitutionprogram*/
#include<stdio.h>
structmacro/*tostoreamacro*/
{
charMname[10];
charMdef[50];
};
FILE*infile,*outfile;
/*AfunctiontobuildMacrotable*/
voidbuild_Mtable(structmacrolist[15],FILE*ptr);
/*functiontoreadawordfrominputtextfile*/
voidget_token(FILE*infile,chartoken[20],char
*break_point,int*flag);
charinputfile[15],outputfile[15];
/*functiontowriteawordtooutputtextfile*/
voidwriteout(FILE*outfile,chartoken[20],charch,
structmacroMlist[15]);
main()

{
chartoken[20];
charbreak_point;
intflag,j;
charch;
structmacroMlist[15];
printf(\nEnterthenameofinputfile:);
fflush(stdin);gets(inputfile);
printf(\nEnterthenameofoutputfile:);
fflush(stdin);gets(outputfile);
infile=fopen(inputfile,r);
outfile=fopen(outputfile,w);
build_Mtable(Mlist,infile);
flag=0;
clrscr();
/*flag==3indicatesEOFencountered*/
while(!feof(infile)&&flag!=3)
{flag=0;
token[0]=\0;
ch=;
get_token(infile,token,&ch,&flag);
writeout(outfile,token,ch,Mlist);
}
fclose(infile);
fclose(outfile);
}/*endofmain*/
voidget_token(FILE*infile,chartoken[30],char
*break_pt,int*flag)
{inti;
charch;
i=1;
while(!(*flag))
{ch=fgetc(infile);if(ch==EOF){*flag=3;
return;}
switch(ch)
{/*breakpointsforwordsortokens*/
case\n:
case:
case,:/*flag==1indicatesthatabreak
pointoccurred*/

case.:
case;:*break_pt=ch;*flag=1;token[++i]=
\0;
break;
default:token[++i]=ch;
}
}
}/*endofget_token*/
voidbuild_Mtable(structmacrolist[15],FILE*ptr)
{chartoken[20],ch1,temp[2];
inti;
token[0]=\0;
i=1;
while(strcmp(token,MEND))
{intflag=0;
charch=;
get_token(ptr,token,&ch,&flag);
if(!strcmp(token,define))
{i++;
flag=0;
ch=;
get_token(ptr,token,&ch,&flag);
strcpy(list[i].Mname,token);
list[i].Mdef[0]=\0;
while(ch!=\n)
{flag=0;
ch=;
get_token(ptr,token,&ch,&flag);
strcat(list[i].Mdef,token);
if(ch!=\n)
{temp[0]=ch;
temp[1]=\0;
strcat(list[i].Mdef,temp);
}
}
}
}
i++;/*Markstheendofthemacrolist*/
strcpy(list[i].Mname,END);
}/*endofbuild_Mtable*/

voidwriteout(FILE*outfile,chartoken[20],charch,
structmacroMlist[15])
{
inti=0;
while(strcmp(Mlist[i].Mname,END))
{
/*replacemacrobyitsdefinition*/
if(!strcmp(Mlist[i].Mname,token))
{fputs(Mlist[i].Mdef,outfile);
fputc(ch,outfile);
return;
};
i++;
}
fputs(token,outfile);
fputc(ch,outfile);
}
Sample Input:
#defineceComputerEngineering
#defineymcaYMCAInstituteofEngineering
#defineaksA.K.Sharma
#definefbdFaridabad
#defineMDUMaharishiDayanandUniversity,Rohtak
MEND
ymca is an Institute situated in fbd. It is a Government of Haryana
institute. It is affiliated to MDU. The institute is rated as the best
engineering college in the state of Haryana. The ce department is the
largest department of the institute. aks works in ce department as a
Professor.
Sample Output:
YMCA Institute of Engineering is an Institute situated in Faridabad.
It is a Government of Haryana institute. It is affiliated to Maharishi
Dayanand University, Rohtak. The institute is rated as the best
engineering college in the state of Haryana. The Computer
Engineering department is the largest department of the institute.
A.K. Sharma works in Computer Engineering department as a
Professor.

Note: The program does not perform syntax checking, i.e., if the terms
define and MEND are misspelled, the program will go hey wire.
Problem 2: Statements in BASIC language are written in such a
manner that each statement starts with an integer number as shown
below:
10REMThisisasamplestatement
20inputA,B
30goto150
:
:
:
100If(xy)goto20
Write a program that reads a file of a BASIC program. For a given offset,
it adds the offset to all the statement numbers. The arguments of the
goto statements should also be modified to keep the program consistent.
In fact, this type of activity is done by a loader, i.e., when it loads a
program into the main memory for execution, it adds the offset of the
address of the main memory to all address modifiable parts of the
program such as to goto or jump statements and the like.
Solution: The get_token()andwriteout() functions of previous
program would be used in this program to read a token and to write the
tokens to a given file. The following functions ofstdlib.h would be
used to convert a string to integer and vice versa.
(1) <int> atoi
(string)

This function takes a string of characters

and converts it into an equivalent integer.

(2) void itoa (int,

string, <radix>)

This function converts an integer to an

equivalent string. Radix is the radix of the
integer such as 2, 10 depending upon the
number being binary or decimal.

The required program is given below:

/*Thisprogramimplementsarelativeloader*/
#include<stdio.h>
#include<stdlib.h>
#include<ctype.h>

/*functiontoreadawordfrominputtextfile*/
voidget_token(FILE*infile,chartoken[20],char
*break_point,int*flag);
charinputfile[15],outputfile[15];
/*functiontowriteawordtooutputtextfile*/
voidwriteout(FILE*outfile,chartoken[20],char
ch);
main()
{
FILE*infile,*outfile;
chartoken[20];
charlast_token[20]=;
charbreak_point;
intflag,j,temp;
intoffset;
charch,ch1;
printf(\nEnterthenameofinputfile:);
fflush(stdin);gets(inputfile);
printf(\nEnterthenameofoutputfile:);
fflush(stdin);gets(outputfile);
printf(\nEnterthevalueofoffset(int):);
scanf(%d,&offset);
infile=fopen(inputfile,r);
outfile=fopen(outputfile,w);
flag=0;
clrscr();
ch1=\n;
/*flag==3indicatesEOFencountered*/
while(!feof(infile)&&flag!=3)
{flag=0;
token[0]=\0;
ch=;
get_token(infile,token,&ch,&flag);
/*Checkifitisanumberappearingatthe
beginningofalineorasargumentof
gotostatement*/
if((isdigit(token[0])&&ch1==\n)||(isdigit
(token[0])&&(!strcmp(last_token,goto))))
{temp=atoi(token);
temp=temp+offset;
itoa(temp,token,10);

}
writeout(outfile,token,ch);
ch1=ch;
strcpy(last_token,token);
}
fclose(infile);
fclose(outfile);
}/*endofmain*/
voidget_token(FILE*infile,chartoken[30],char
*break_pt,int*flag)
{inti;
charch;
i=1;
while(!(*flag))
{ch=fgetc(infile);if(ch==EOF){*flag=3;
return;}
switch(ch)
{/*breakpointsforwordsortokens*/
case\n:
case:
case,:/*flag==1indicatesthatabreakpoint
occurred*/
case.:
case;:*break_pt=ch;*flag=1;token[++i]=
\0;
break;
default:token[++i]=ch;
}
}
}/*endofget_token*/
voidwriteout(FILE*outfile,chartoken[20],charch)
{
fputs(token,outfile);
fputc(ch,outfile);
}
Sample input:
10Remthisisaloader
20InputA,B
30If(A20)goto20

40C5A15*20;
50if(C50)goto100
60C5C1B
70goto50
100stop
Valueofoffset550
Sample output:
60Remthisisaloader
70InputA,B
80If(A20)goto70
90C5A15*20;
100if(C50)goto150
110C5C1B
120goto50
150stop
EXERCISES
1.
2.
3.
4.
5.

What are the different methods of opening a file? Explain in brief.

Differentiate between fwrite()andfput() functions.
How can the end of a file be detected? Explain in brief.
Write a program which counts the number of record in a given file.
Give suitable declarations for the following files:

a. A file containing components where each component

consists of the following information:
Name:

20 characters

Code:

integer type

NET:

float type

b. A text file called Book.

2. Write a program which creates an exact copy of a given file called Ofile in
the form of a new file called Nfile. Choose appropriate component type for
this problem.
3. Write a program which counts the number of times a given alphabet stored
in variable ch appears in a text file.

4. Write a program which searches a given file for a record equal to an input
value of emp-code. The component record structure of the file is as given
below:

Choose appropriate data types wherever necessary.

5. Write a program which reads a text file called document and generates a
table of frequency count of alphabets appearing in the text. The output
should be displayed in the following form: Frequency count
Alphabet

Count

A
B
C
.
.
.
Z
6. A student record is defined as given in Example 5. Write a program which
reads a file of such records (say class-file) and prints the result in the
following format:
S.No.

Condition

No.ofStudents

Marks<50%

xx

Marks>50%and<60%

xx

Marks>60%and<75%

xx

Marks>75%and<90%

xx

Marks>=90

xx

7. (Assume Max marks in each subject = 100).

8. What is EOF? Explain its utility.
9. Modify Example 11 such that it generates a list of students who secure more
than 80 per cent marks.
10. Write an interactive menu driven C program to create a text file and then
display the file. Create another text file by converting each line of the newly
created text file into a lowercase string. Display the newly created file. In
case number of lines exceeds 22, file should be displayed one screen at a
time.
11. Write an interactive menu driven C program to create a text file and then
display the file. Create another text file by reversing each line of the newly

created text file. Display the newly created file. In case number of lines
exceeds 22, file should be displayed one screen at a time.
12. Explain relative files in detail.
13. Explain indexed sequential organization.
14. Write an explanatory note on multilevel indexed files.
15. Define file activity ratio, response time, and volatility.
16. How records are edited or deleted from an indexed file?

10 Chapter
Advanced Data Structures
CHAPTER OUTLINE

10.1 AVL Trees

10.2 Sets
10.3 Skip Lists
10.4 B-Trees
10.5 Search by Hashing
10.1 AVL TREES

During the discussion on binary search trees (BST) in Chapter 7, it was

appreciated that BST is an excellent tool for fast search operations. It
was also highlighted that if the data input to BST is not properly
planned, then there is every likelihood that the BST may end up in a
skewed binary tree or at least a lopsided binary tree wherein one subtree
has more height than the other subtree. Consider the data given below:

Cat Family 5 {Lion, Tiger, Puma, Cheetah, Jaguar, Leopard, Cat,

Panther, Cougar, Lynx, Ocelot}
This data can be provided in any combination and the
resultant BSTs would be different from each other depending upon the
pattern of input. Some of the possible BSTs produced from the above
data are shown in Figure 10.1.

Fig. 10.1 BSTs produced from same data provided in different

combinations
It may be noted that the BST of Figure 10.1 (a) is a skewed binary tree
and the search within such a data structure would amount to a linear
search. The BST of Figure 10.1 (b) is heavily lopsided towards left
whereas the BST of Figure 10.1 (c) is somewhat balanced and may lead to
a comparatively faster search. If the data is better planned, then
definitely an almost balanced BST can be obtained. But the planning can
only be done when the data is already with us and is suitably arranged
before creating the BST out of it. However, in many applications such as
symbolprocessing in compilers, the data is dynamic and
unpredictable and, therefore, creation of a balanced BST is a remote
possibility. Hence, the need to develop a technique that maintains a
balance between the heights of left and right subtrees of a binary tree has
always been felt; the major aim being that whatever may be the order of
insertion of nodes, the balance between the subtrees is maintained.
A height balanced binary tree was developed by Adelson-Velenskii and
Landis, Russian researchers, in 1962 which is popularly known as an
AVL Tree. The AVLTree assumes the following basic subdefinitions:
1. The height of a tree is defined as the length of the longest path from the root
node of the tree to one of its leaf nodes.
2. The balance factor (BF) is defined as given below:

BF = height of left subtree (HL) height of right subtree (HR)

With the above subdefinitions, the AVLTree is defined as a balanced

binary search tree if all its nodes have balance factor (BF) equal to 1, 0,
or 1.
In simple words we can say that in AVLTree, the height of two subtrees
of a node differs by at most one. Consider the trees given in Figure 10.2.

Fig. 10.2 The binary search trees with BF labelled at each node
It may be noted that in Figure 10.2 (a), all nodes have BF within the
range, i.e. 1, 0, or 1. Therefore, it is an AVLTree. However, the root of
graph in Figure 10.2 (b) has BF equal to 2, which is a violation of the
rule and hence the tree is not AVL.
Note: A complete binary search tree is always height balanced but a
height balanced tree may or may not be a complete binary tree.
The height of a binary tree can be computed by the following simple
steps:
1. If a child is NULL then its height is 0.
2. The height of a tree is = 1 + max (height of left subtree, height of right
subtree)

Based on the above steps, the following algorithm has been developed to
compute the height of a general binary tree:
Algorithm compHeight(Tree)
{
if(Tree==NULL)
{height=0;returnheight}
hl=compHeight(leftChild(Tree));
hr=compHeight(rightChild(Tree));
if(hl>=hr)height=hl+1;
else
height=hr+1;
returnheight;
}

The following operations can be performed on an AVLTree:

1. Searching an AVLTree.
2. Inserting a node in an AVLTree.
3. Deleting a node from an AVLTree.

10.1.1 Searching an AVLTree

An AVLTree can be searched for a key K by visiting nodes starting from
root. The visited node is checked for the presence of K. If found, then the
search stops otherwise if value of K is less than the key value of visited
node, then the left subtree is searched else the right subtree is searched.
The algorithm for AVL search is given below:
Algorithm searchAVL(Tree,K)
{
if(DATA(Tree)==K)
{promptSearchsuccessful;Stop;}
if(DATA(Tree)<K)
searchAVL(leftChild(Tree),K);
else
searchAVL(rightChild(Tree),K);
}
The search for key (K = 37) will follow the path as given in Figure 10.3.

Fig. 10.3 Search for a key within an AVLTree

10.1.2 Inserting a node in an AVL Tree
Whenever a node is inserted into a subtree which is full and whose
height is already 1 more than its sibling, then the subtree becomes
unbalanced. This imbalance in a subtree ultimately leads to imbalance of
whole of the tree as shown in Figure 10.4. An insertion of node 50 has
caused an imbalance in the tree. The nearest ancestor whose BF has
become 2 is called a pivot. The node 30 is the pivot in the unbalanced
tree of Figure 10.4.

Fig. 10.4 An insertion into a full subtree

The imbalance can be removed by carefully rearranging the nodes of the
tree. For instance, the zone of imbalance in the tree is the left subtree
with root node 30 (say Node P). Within this subtree, the imbalance is in
right subtree with root node 46 (say Node Q). The closest node to the
inserted node is 57. Call this node as R. A point worth noting is that P, Q,
and R are numbers and can be rearranged as per binary search rule, i.e.,
the smallest becomes the left child, the largest as the right child and the
middle one becomes the root. This rearrangement can eliminate the
imbalance in the section as shown in Figure 10.5.

Fig. 10.5 The rearrangement of nodes

It may be noted that imbalanced section of the tree has been rotated to
left to balance the subtree. The rotation of a subtree in an AVLTree is
called an AVLrotation. Depending upon, in which side of the pivot a
node is inserted, there are four possible types of AVL rotations as
discussed in following sections.
10.1.2.1 Insertion of a Node in the Left Subtree of the Left Child of the Pivot
In this case, when a node in an AVLTree is inserted in the left subtree
of the left child (say LL) of the pivot then the imbalance occurs as shown
in Figure 10.6 (a).
It may be noted that in this case, after the rotation, the pivot P has
become the right child, Q has become the root. QR, the right child of Q
has become the left child of P. The QL, the left child of Q remains intact.

Fig. 10.6 AVL rotation towards right

The algorithm for this rotation is given below:
/*TheTreepointstothePivot*/
Algorithm rotateLL(Tree)
{
ptr=leftChild(Tree);/*pointptrtoQ*/
nptr=rightChild(ptr);/*pointnptrtoright
childofQ*/
rightChild(ptr)=Tree;/*attachPasrightchild
ofQ*/
leftChild(Tree)=nptr;/*attachrightchildofQ
asleftchildofP*/
Tree=ptr;/*Qbecomestheroot*/
returnTree;
}
Consider the AVLTree given in Figure 10.7 (a). The tree has become
imbalanced after the insertion of 12 [see Figure 10.7 (b)].

Fig. 10.7 The AVL rotation (LL)

Figure10.7 (c) shows the application of pointers on the nodes of the tree
as per the AVL rotation.Figure 10.7 (d) shows the final AVLTree (height
balanced) after the AVL rotation, i.e., after the rearrangement of nodes.
10.1.2.2 Insertion of a Node in the Right Subtree of the Right Child of the Pivot
In this case, when a node in an AVLTree is inserted in the right subtree
of the right child (say RR) of the pivot then the imbalance occurs as
shown in Figure 10.8 (a).

Fig. 10.8 AVL rotation towards left

It may be noted that in this case, after the rotation, the pivot P has
become the left child, Q has become the root. QL, the left child of Q (i.e,
QL) has become the right child of P. The QR, the right child of Q remains
intact.
The algorithm for this rotation is given below:
/*TheTreepointstothePivot*/
Algorithm rotateRR(Tree)
{
ptr=rightChild(Tree);/*pointptrtoQ*/
nptr=leftChild(ptr);/*pointnptrtoleftchild
ofQ*/
leftChild(ptr)=Tree;/*attachPasleftchildof
Q*/
rightChild(Tree)=nptr;/*attachleftchildofQ
asrightchildofP*/
Tree=ptr;/*Qbecomestheroot*/
returnTree;
}
Consider the AVLTree given in Figure 10.9 (a). The tree has become
imbalanced after the insertion of 100 (see Figure 10.9 (b)).
Figure 10.9 (c) shows the application of pointers on the nodes of the tree
as per the AVL rotation.Figure 10.9 (d) shows the final AVL

Tree (height balanced) after the AVL rotation, i.e., after the
rearrangement of nodes.
10.1.2.3 Insertion of a Node in the Right Subtree of the Left Child of the Pivot
In this case, when a node in an AVLTree is inserted in the right subtree
of the left child (say LR) of the pivot then the imbalance occurs as shown
in Figure 10.10 (a).

Fig. 10.9 The AVL rotation (RR)

Fig. 10.10 The double rotation Left-Right (LR)

This case requires two rotations: rotation 1 and rotation 2.
Rotation 1: In this rotation, the left child of R becomes the right child
of Q and R becomes the left child of P. Q becomes the left child of R. The
right child of R remains intact.
Rotation 2: In this rotation, the right child of R becomes the left child
of P. R becomes the root. P becomes the right child of R.

The final balance tree after the double rotation is shown in Figure 10.10
(b).
The algorithm for this double rotation is given below:
/*TheTreepointstothePivot*/
Algorithm rotateLR(Tree)
{
ptr=leftChild(Tree);/*pointptrtoQ*/
rotateRR(ptr);/*performthefirstrotation*/
ptr=leftChild(Tree);/*pointptrtoR*/
rotateLL(ptr);
}
The trace of the above algorithm is shown in Figure 10.11 wherein
the AVLTree has become imbalanced because of insertion of node 50
in the right subtree of left child of pivot [see Figures 10.11 (a) and (b)].

Fig. 10.11 The trace of double rotation on an imbalanced AVLTree

After the double rotation, the tree has again become height balanced as
shown in Figures 10.11 (c) and(d).
10.1.2.4 Insertion of a Node in the Left Subtree of the Right Child of the Pivot
In this case, when a node in an AVLTree is inserted in the left subtree
of the right child (say LR) of the pivot then the imbalance occurs as
shown in Figure 10.12 (a).
This case also requires two rotations: rotation 1 and rotation 2.

Rotation 1: In this rotation, the right child of R becomes the left child
of Q and R becomes the right child of P. Q becomes the right child of R.
The left child of R remains intact.
Rotation 2: In this rotation, the left child of R becomes the right child
of P. R becomes the root. P becomes the left child of R.
The final balance tree after the double rotation is shown in Figure 10.12
(b).
Following is the algorithm for this double rotation:
/*TheTreepointstothePivot*/
Algorithm rotateRL(Tree)
{
ptr=rightChild(Tree);/*pointptrtoQ*/
rotateLL(ptr);/*performthefirstrotation*/
ptr=rightChild(Tree);/*pointptrtoR*/
rotateRR(ptr);
}

Fig. 10.12 The double rotation Right-Left (RL)

The trace of the above algorithm is shown in Fig. 10.13 wherein the AVL
Tree has become imbalanced because of insertion of node 92 the left
subtree of right child of pivot (see Fig. 10.13(a), (b)). After the double
rotation, the tree has again become height balanced as shown in Fig.
10.13(c),(d).

Fig. 10.13 The trace of double rotation on an imbalanced AVLTree

10.2 SETS

A set is an unordered collection of homogeneous elements. Each element

occurs at most once with the following associated properties:
1. All elements belong to the Universe. The Universe is defined as all potential
elements of set.
2. An element is either a member of the set or not.
3. The elements are unordered.

Some examples of sets are given below.

Cat_ Family = {Lion, Tiger, Puma, Cheetah, Jaguar, Leopard, Cat,
Panther, Cougar, Lynx, Ocelot}
Fruits = {mango, orange, banana, grapes, apple, guava, peach}
LAN = {Comp, Mech, Elect, Civil, Accts, Estb, Mngmt, Hostel}
Note: The LAN is a set of nodes of a local area network of an educational
institute spread over various departments like computer engineering,
mechanical engineering, electrical engineering, etc.
The basic terminology associated with sets is given below:
1. Null set: If a set does not contain any element, then the set is called empty
or null set. It is represented by .
2. Subset: If all elements of a set S1 are contained in set S2, then S1 is called a
subset of S2. It is denoted by S1 S2.
Example: The set of big_Cat = {Lion, Tiger, Panther, Cheetah, Jaguar} is a
subset of the set ofCat_Family.

3. Union of sets: The union of two sets S1 and S2 is obtained by collecting all
members of either set without duplicate entries. This is represented
by S1 S2.
Example: If S1 = {1, 3, 4}, S2 5 {4, 5, 6, 7, 8}, then S1 S2 = {1, 3, 4, 5, 6,
7, 8}.
4. Intersection of sets: The intersection of two sets S1 and S2 is obtained by
collecting common elements of S1 and S2. This is represented by S1 S2.
Example: If S1 = {1, 3, 4}, S2 = {4, 5, 6, 7, 8}, then S1 S2 = {4}.
5. Disjoint sets: If two sets S1 and S2 do not have any common elements,
then the sets are called disjoint sets, i.e., S1 S2 = .
Example: S1 = {1, 3}, S2 = {4, 5, 6, 7, 8} are disjoint sets.
6. Cardinality: Cardinality of a set is defined as the number of unique
elements present in the set. The cardinality of a set S is represented as |S|.
Example: If S1 = {1, 3, 4}, S2 = {4, 5, 6, 7, 8}, then |S1| = 3 and |S2| = 5.
7. Equality of sets: If S1 is subset of S2 and S2 is subset of S1, then S1 and S2
are equal sets. This means that all elements of S1 are present in S2 and all
elements of S2 are present in S1.The equality is represented as S1 S2
Example: If S1 = {1, 3, 4}, S2 = {3, 1, 4}, then S1 S2.

Fig. 10.14 Various set operations

8. Difference of sets: Given two sets S1 and S2, the difference S1 S2 is
defined as the set having all the elements which are in S1 but not in S2.
Example: If S1 = {1, 3, 4}, S2 = {3, 2, 5}, then S1 S2 = {1, 4}.
The graphic representation of various operations related to sets is shown
in Figure 10.14.
9. Partition: It is a collection of disjoint sets belonging to same Universe.
Thus, the union of all the disjoint sets of a partition would comprise the
Universe.

10.2.1 Representation of Sets

A set can be represented by the following methods:
1.
2.
3.
4.

List representation
Hash table representation.
Bit vector representation
Tree representation

List and hash table representations are more popular methods as

compared to other methods. A brief discussion on list and hash table
representations is given in following sections.
10.2.1.1 List Representation
A set can be very comfortably represented using a linear linked list. For
example, the set of fruits can be represented as shown in Figure 10.15.

Fig. 10.15 The list representation of a set

Since the set is an unordered collection of elements, linked list is the
most suitable representation for the sets because elements can be
dynamically added or deleted from a set.
10.2.1.2 Hash Table Representation
Hash table representation of a set is also a version of linked
representation. A hash table consists of storage locations called buckets.
The size of a bucket is arbitrary, i.e., it can hold any number of elements
as it stores them into a linked list.Consider the following set:
Tokens = {126, 235, 100, 317, 68, 129, 39, 423, 561, 222, 986}
Let us store the above set in a hash table of five buckets with hash
function as given below:
H(K) = K mod 5
Where, K is an element;
H(K) gives the bucket number in which the element K is to be stored.
The arrangement is shown in Figure 10.16.
Tokens = {126, 235, 100, 317, 68, 129, 39, 423, 561, 222, 986}

Fig. 10.16 Hash table representation of a set

It may be noted that each bucket is having a linked list consisting of
nodes containing a group of elements from the set.
10.2.2 Operations on Sets
The following operations can be defined on sets:
1.
2.
3.
4.

Union
Intersection
Difference
Equality

A brief discussion on each operation is given in the sections that follow.

10.2.2.1 Union of Sets
As per the definition, given two sets S1 and S2, the union of two sets is
obtained by collecting all the members of either set without duplicate
entries. Let us assume that the sets have been represented using lists as
shown in Figure 10.17. The union set S3 of S1 and S2 has been obtained
by the following steps:
1. Copy the list S1 to S3.
2. For each element of S2, check if it is member of S1 or not and if it is not
present in S1 then attach it at the end of S3.

Fig. 10.17 Union of sets

The algorithm for union of two sets S1 and S2 is given below:
/*Thisalgorithmusestwofunctions:copy()and
ifMember().Thecopy()functionisusedtocopya
settoanotheremptyset(sayS1toS3).TheifMember
functionisemployedtotestthemembershipofan
elementinagivenset*/
Algorithm unionOfSets(S1,S2)
{
S3=Null;
S3=copy(S1,S3,last);/*copySetS1toS3*/
ptr=S2;
while(ptr!=Null)
{
Res=ifMember(ptr,S1);/*Checkmembershipof
ptrinS1*/
if(Res==0)
{/*ifthenodeofS2isnotmemberofS1*/
nptr=newnode;
DATA(nptr)=DATA(ptr)
NEXT(nptr)=NULL;
NEXT(last)=nptr;
last=nptr;
}
ptr=Next(ptr);
}
returnS3;
}

/*Thisfunctionisusedtocopyasettoanother
emptyset,i.e.,Set1toSet3*/
Algorithm copy(S1,S3)
{
ptr=S1;
back=S3;
ahead=Null;
while(ptr!=Null)
{ahead=newnode;
DATA(ahead)=DATA(ptr);
NEXT(ahead)=Null;
if(back==Null)/*firstentryintoS3i.e.S3is
alsoNull*/
{
back=ahead;
S3=ahead;
}
else
{
NEXT(back)=ahead;
back=ahead;
}
ptr=NEXT(ptr);
}
returnS3;
}
/*Thisfunctionisemployedtotestthemembership
ofanelementpointedbyptr,inagivensetS*/
Algorithm ifMember(ptr,S)
{
Flag=0;
nptr=S;
while(nptr!=Null)
{
if(DATA(ptr)==DATA(nptr))
{
Flag=1;
break;
}

nptr=NEXT(nptr);
}
returnFlag;
}
10.2.2.2 Intersection of Sets
As per the definition, given two sets S1 and S2, the intersection of two
sets is obtained by collecting all the common members of both the sets.
Lets assume that the sets have been represented using lists as shown
in Figure 10.18. The intersection set S3 of S1 and S2 has been obtained
by the following steps:
1. Initialize S3 to Null;
2. last = S3;
3. For each element of S2, check if it is member of S1 or not and if it is present
in S1 then attach it at the end of S3.

Fig. 10.18 Intersection of sets

The algorithm for intersection of two sets S1 and S2 is given below:
/*ThisalgorithmusesthefunctionifMember()which
teststhemembershipofanelementinagivenset*/
Algorithm intersectionOfSets(S1,S2)
{
S3=Null;
last=S3;
ptr=S2;
while(ptr!=Null)
{
Res=ifMember(ptr,S1);/*Checkmembershipof
ptrinS1*/
if(Res==1)

{/*ifthenodeofS2ismemberofS1*/
nptr=newnode;
DATA(nptr)=DATA(ptr)
NEXT(nptr)=NULL;
If(last==Null)/*S3isemptyanditisfirst
entry*/
{
last=nptr;
S3=last;
}
else
{
NEXT(last)=nptr;
last=nptr;
}
}
ptr=Next(ptr);
}
returnS3;
}
Note: The ifMember() function is already defined in the previous
section.
10.2.2.3 Difference of Sets
As per the definition, given two sets S1 and S2, the difference S1 S2 is
obtained by collecting all the elements which are in S1 but not in S2. Let
us assume that the sets have been represented using lists as shown
in Figure 10.19. The difference set S3, i.e., S1 - S2 has been obtained by
the following steps:
1. Initialize S3 to Null;
2. last =Null;
3. For each element of S1, check if it is a member of S2 or not and if it is not
present in S2 then attach it at the end of S3.

Fig. 10.19 Differences of sets

The algorithm for difference of two sets S1 and S2 is given below:
/*ThisalgorithmusesthefunctionifMember()which
teststhemembershipofanelementinagivenset*/
Algorithm differenceOfSets(S1,S2)
{
S3=Null;
last=S3;
ptr=S1;
while(ptr!=Null)
{
Res=ifMember(ptr,S2);/*Checkmembershipof
ptrinS2*/
if(Res==0)
{/*ifthenodeofS1isnotmemberofS2*/
nptr=newnode;
DATA(nptr)=DATA(ptr)
NEXT(nptr)=NULL;
If(last==Null)/*S3isemptyanditisfirst
entry*/
{
last=nptr;
S3=last;
}
else
{
NEXT(last)=nptr;
last=nptr;
}
}

ptr=Next(ptr);
}
returnS3;
}
Note: The ifMember() function is already defined in the previous
section.
10.2.2.4 Equality of Sets
As per the definition, given two sets S1 and S2, if S1 is subset of S2 and
S2 is subset of S1 then S1 and S2 are equal sets. Let us assume that the
sets have been represented using lists. The equality S1 = S2 has been
obtained by the following steps:
1. For each element of S1, check if it is a member of S2 or not and if it is not
present in S2 then report failure.
2. For each element of S2, check if it is a member of S1 or not and if it is not
present in S1 then report failure.
3. If no failure is reported in step1 and step 2, then infer that S1 = S2.

The algorithm for equality of two sets S1 and S2 is given in the following:
/*ThisalgorithmusesthefunctionifMember()which
teststhemembershipofanelementinagivenset*/
Algorithm equalSets(S1,S2)
{
ptr=S1;
flag=1;
while(ptr!=Null)
{
Res=ifMember(ptr,S2);/*Checkmembershipof
ptrinS2*/
if(Res==0)
{flag=0;
break;
}
ptr=NEXT(ptr);
}
if(flag==0){promptnotequal;exit();}
ptr=S2;
while(ptr!=Null)
{

Res=ifMember(ptr,S1);/*Checkmembershipof
ptrinS1*/
if(Res==0)
{flag=0;
break;
}
ptr=NEXT(ptr);
}
if(flag==0){promptnotequal;exit();}
else
promptSetsEqual;
}
Note: The ifMember() function is already defined in the previous
section.
10.2.3 Applications of Sets
There are many applications of sets and a few of them are given below:
(1) Web linked pages: For each item displayed to user, Google offers a
link called Similar pages (seeFigure 10.20), which is in fact the list of
URLs related to that item. This list can be maintained as a set of related
pages by using Union operation.

Fig. 10.20 The link to similar URLs

Consider the web pages visited by a user as shown in Figure 10.21.
The arrows between the various pages indicate the link between them.
Initially, all pages are P9 stored as singletons in the order of their visit by
P7 the users as given below:
{P4}, {P11}, {P7}, {P6}, {P5}
Afterwards, the related pages are grouped together by union of the
participating singletons as shown below:
1. {P4} {P7} {P11} {P9}
2. {P6} { P12} { P4}
3. {P5}

We get the following three related partitions, i.e., the disjoint sets:
1. P4, P7, P11, P9
2. P6 P12 P4
3. P5

From the above representation, it can be found as to whether two

elements belong to the same set or not. For example, P6 and P12 belong
to the same set. A click on page P11 would produce a list of links to P4, P7
and P9, grouped as Similar pages to P11.

Fig. 10.21 The web pages visited by a user

(2) Colouring black and white photographs: This is an interesting
application in which an old black and white photograph is converted into
a coloured one. The method is given below:
1. The black and white photograph is digitized, i.e., it is divided into pixels of
varying shades of grey, i.e., ranging from white to black.
2. Pixels belonging to the same part or component are called equivalent pixels.
For example, pixels belonging to eyeballs are equivalent. Similarly, pixels
belonging to the shirt of a person are equivalent.
3. The equivalence is decided by comparing the grey levels of adjacent pixels.
4. In fact, 4 pixels are taken at a time to find the equivalent pixels among the
adjacent pixels. This activity is also called 4-pixel square scan.
5. The scanning starts from left to right and top to bottom.
6. Each pixel is assigned a colour.
7. The equivalent pixels are given the same colour.
8. If two portions with different coloured pixels meet each other, then it is
considered as part of the same component and then both the portions are
coloured with the same colour chosen from either of the colours as shown
in Figure 10. 22.

Fig. 10.22 The colouring of black and white photographs

It may be noted that the equivalence relation between the pixels is
obtained through set operations where an equivalence relation R over a
set S can be seen as a partitioning of S into disjoint sets.
(3) Spelling checker: A spelling checker for a document editor is
another interesting application of sets, especially the hash table
representation. A normal dictionary is maintained using hash table
representation of sets. A simple-most dictionary (Dict.) will have 26
buckets because there are 26 alphabets in English language. Similarly,
the words of a document (Doc) would also be represented in the same
fashion as shown in Figure 10.23. A difference operation Doc Dict would produce all those words which are in document but not in
dictionary, i.e., the misspelled words.

Fig. 10.23 The working of spelling checker

10.3 SKIP LISTS

Search is a very widely used operation on lists. A one-dimensional array

is most suitable data structure for a list that does not involve insertion
and deletion operations. If the list is unsorted, linear search is used and

for a sorted list, binary search is efficient and takes O(log

n) operations.
For random input, AVLTrees can be used to search, insert, and delete
numbers with time complexity as O(logn). However, the balancing
of AVLTrees requires rotations in terms of extra o(logn)operations.
Binary search trees tend to be skewed for random input.
However, when insertion and deletion operations are important and the
search is also mandatory then a list can be represented using a sorted
linked list. A search operation on a sorted linked list takes o(n)
operations.
Now, the question is how to make efficient search operation on sorted
linked lists? In 1990, Bill Pugh proposed an enhancement on linked lists
and the new data structure was termed as skip list.
A skip list is basically a sorted linked list in ascending order. Now, extra
links are added to selected nodes of the linked list so that some nodes
can be skipped while searching within the linked list so that the overall
search time is reduced. Consider the normal linked list shown in Figure
10.24. Let us call this list as level 0.

Fig. 10.24 A sorted linked list

It may be noted that the time complexity of various operations such
as search, insertion, deletion, etc. on a sorted linked list is o(N).
Let us now add extra links in such a manner that every alternative node
is skipped. The arrangement is shown in Figure 10.25. The chain of extra
links is called level 1. The head node is said to have height equal to 1.

Fig. 10.25 A skip list with level 1

Now, we carry on to add another level of links, i.e., level 2 wherein every
alternate node of level 1 is skipped as shown in Figure 10.26. The height
of head node has become 2.
The levels of links are added till we reach a stage where the link from
head node has reached almost to the middle element of the linked list as
has happened in Figure 10.26.

Fig. 10.26 A skip list with level 2

It may be noted that in level 1, links point to every second element in the
list. In level 2, the links point to every fourth element in the list. Thus, it
can be deduced that a link in the ith level will point to 2*ith element in
the list.A point worth noting is that the last node of the skip list
is pointed by links of all levels.
Now the search within the skip list shall take O(log n) time as it will
follow binary search pattern. For example, the search for 53 will follow
the route shown in Figure 10.27.

Fig. 10.27 The path followed for searching 53 in the list

The algorithm that searches a value VAL in the skip list is given below:
Algorithm skipListSearch()
{
curPtr=Head;
For(I=N;I>1;I)
{
while(DATA(curPtr[i]<VAL)
CurPtr[I]=NEXT(CurPtr[I]);
If(DATA(curPtr[I]))==VAL);returnsuccess;
}
returnfailure;
}
It may be noted that insertion and deletion operations in the list will
disturb the balance of the skip list and many levels of pointer adjustment
would be required to redistribute the levels of pointers and heights of
head and intermediate nodes.

10.4 B-TREES

A binary search tree (BST) is an extremely useful data structure as far as

search operations on static data items is concerned. The reason is that
static data can be so arranged that the BST generated out of the data is
almost height balanced. However, for very large data stored in files, this
data structure becomes unsuitable in terms of number of disk accesses
required to read the data from the disk. The remedy is that each node
of BST should store more than one record to accommodate the complete
data of the file. Thus, the structure of BST needs to be modified. In fact,
there is a need to extend the concept of BST so that huge amount of data
for search operations could be easily handled.
We can look upon a BST as a two-way search tree in which each node
(called root) has two subtreesleft subtree and right subtree, with
following properties:
1. The key values of the nodes of the left subtree are always less than the key
value of the root node.
2. The key values of the nodes of the right subtree are always more than the
key value of the root node.

The concept of two-way search tree can be extended to create an m

way search tree having node structure shown in Figure 10.28.
The mway tree has following properties:
1. Each node has any number of children from 2 to M, i.e., all nodes
have degree<=M, where M>=2.
2. Each node has keys (K1 to Kn) and pointers to its children (Po to Pn), i.e.,
number of keys is one less than the number of pointers. The keys are
ordered, i.e., Ki < Ki+1for1<=i<n.
3. The subtree pointed by a pointer Pi has key values less than the key value of
Ki+1 for 1 <= i < n.
4. The subtree pointed by a pointer Pn has key values greater than the key
value of Kn.
5. All subtrees pointed by pointers Pi are mwaytrees.

Fig. 10.28 M-way tree

It may be noted that all keys in the subtree to the left of a key are
predecessors of the key and that on the right are successors of the key as

shown in Figure 10.29. The mway tree shown in the figure is of the
order 3, i.e.,m = 3.
A close look at the mway tree suggests that it is a multilevel index
structure. In order to have efficient search within an m-way tree, it
should be height balanced. In fact,a height balanced mwaytreeis
calledasBtree. A precise definition of a Btree is given below
1. The root has at least two children. If the tree contains only a root, i.e., it is a
leaf node then it has no children.
2. An interior node has between |m/2| and m children.
3. Each leaf node must contain at least |m/2| 1 children.
4. All paths from the root to leaves have the same length, i.e., all leaves are at
the same level making it height balanced.

Fig. 10.29 An m-way search tree

Consider the Btree given in Figure10.30. It is of order 5 because all
the internal nodes have at least |5/2| = 3 children and two keys. In fact,
the maximum number of children and keys are 5 and 4, respectively. As
per rule number 3, each leaf node must contain at least |5/2| 1 = 2
keys.

Fig. 10.30 A B-tree of size 5

It may be noted that the node of a Btree, by its nature, can
accommodate more than one key. We may design the size of a node to be
as large as a block on the disk (say a sector). As a block is read at a time
from the disk, compared to a normal BST, less number of disk accesses
would be required to read/ write data from the disk.
The following operations can be carried out on a Btree:
1. Searching a Btree
2. Inserting a new key in Btree
3. Deleting a key from a Btree

10.4.1 Searching a Key in a B-Tree

A key K can be searched in a Btree in a fashion similar to a BST. The
search within a node, including root, follows the following algorithm:
Algorithm searchBTree()
{
Step
1.CompareKwithKifor1<=i<=n.
2.IfKi==K,thenreportsearchsuccessfuland
return.
3.FindthelocationisuchthatKi<=K<K<i+1
4.MovetothechildnodepointedbyPi.
5.IfPiisequaltoNULL,thenreportfailureand
returnelserepeatsteps1to4.
}
The search for K = 82 will follow the path as shown in Figure 10.31.

Fig. 10.31 Search within a B-tree

10.4.2 inserting a Key in a B-Tree

The insertion in a Btree involves three primary steps: search - insert (optional) split a node. The search operation finds the proper node
wherein the key is to be inserted. The key is inserted in the node if it is
already not full, else the node is split on the median (middle) key and the
median number is inserted in the parent node. An algorithm that inserts
a key in a Btree is as follows:
Algorithm insertKey(Key)
{
Step
1.SearchthekeyintheBtree.
2.IfitisfoundthenpromptDuplicateKey.
3.Ifnotfound,thenmarktheleaf./*
Unsuccessfulsearchalwaysendsataleaf*/
4.Iftheleafhasemptylocationforthekey,then
insertitattheproperplace.
5.Iftheleafisfull,thensplittheleafinto
twohalvesatthemediankey.
6.Movethemediankeyfromtheleaftoitsparent
nodeandinsertitthere.
7.Insertthekeyintotheappropriatehalf.
8.Adjustthepointersoftheparentsothatthey
pointtothecorrecthalf.
9.STOP
}
Let us insert the following list of numbers into an empty Btree:
3, 14, 7, 1, 8, 5, 11, 17, 13, 6, 23, 12, 20, 26
The trace of the above algorithm on the list of numbers is given
in Figures 10.32 and 10.33.

Fig. 10.32 Insertion of keys in a B-tree

10.4.3 Deleting a Key from a B-Tree
Deletion of a key from a node of Btree has to be done with lot of care
because it might violate the basic properties of the Btree itself. For
example, one must consider the following cases:
1. After the deletion of a key in a leaf node, if the degree of the node decreases
below the minimum degree of the tree then the siblings must be combined
to increase the degree to the acceptable level.
Consider the deletion of K=8 from the Btree given in Figure 10.34.
Since the key (8) is in the leaf of the Btree and its deletion does not
violate the degree of that leaf, the key is deleted and the other elements in
the node are appropriately shifted to maintain the order of the keys.
2. If the node is internal and has children, the children must be rearranged so
that the basic properties of the Btree are conformed to. In fact, we find its
successor from the leaf node. If key is K and its successor is S, then S is
deleted from the leaf node. Finally, S replaces the K.

Fig. 10.33 Insertion of keys in a B-tree

Consider the deletion of 20 from the Btree of Figure 10.35.
It may be noted that 23, the successor of 20, has been deleted from the
child of node containing 20. The successor (23) has replaced the key (20)
as the final step.

Fig. 10.34 Deletion of a key from a leaf of the B-tree

Consider the deletion of 19 from the Btree of Figure 10.36.
It may be noted that although the node containing 19 is a leaf node, the
deletion violates the property that there should be at least two keys in the
leaf node of a Btree of order 5. Therefore, if an immediate sibling (left
or right) has an extra key, then a key is borrowed from the parent (23)
that is stored in place of the deleted key (19). The extra key (24) from the
sibling is shifted to the parent as shown in Figure 10.36.
Let us finally consider the case of deletion of key (5). The deletion of 5
leaves no extra key in the leaf and its siblings (left or right) have also no
extra keys to donate. Therefore, the parent key (4) is moved down and
the siblings are combined as shown in Figure 10.37(a).
The moving down of 4 has caused a new problem that the parent node
now contains only one key 7 (seeFigure 10.37), which is another
violation. As its sibling is also not having an extra key, the siblings need
to be combined and bring 13 from the parent node. The final tree after
adjustment is shown in Figure 10.37(b). However, the height of the tree
has reduced by 1.

Fig. 10.35 Deletion of a key from an internal node of the B-tree

Note: From the above discussion, it is clear that the insertion, and
especially deletion, is a complicated operation in Btrees. Therefore,
the size of leaf nodes is kept as large as 256 or more so that insertion and
deletion do not demand rearrangement of nodes.
10.4.4 Advantages of B-Trees
1. As the size of a node of a Btree is kept as large as 256 or more, the height
of the tree becomes very small. Such a node supports children in the range
of (128 to 256). Therefore, insertion and deletion operations will generally
remain within the limit without causing over-or underflow of the node.
2. As the height of a Btree is very small, the key being searched is normally
found at a shallow depth.
3. If the root and next level of nodes are kept in the main memory (RAM),
more than a million records of file can be indexed. Consequently, fewer disk
accesses would be required to read the required record from the disk.

Fig. 10.36 Deletion of a key from a leaf causing violation

Fig. 10.37 (a) B-tree after deletion of key, k = 5

Fig. 10.37 (b) B-tree after double violation

10.5 SEARCHING BY HASHING

In system programming,search is a very widely used operation. An

assembler or a compiler very frequently searches the presence of a token
in a table such as symbol table, opcode table, literal tableetc. Earlier in
the book, following techniques have been discussed for search
operations:
1. Linear search
2. Binary search
3. Binary search trees

Averagely, the linear search requires O(n) comparisons to search an

element in a list. On the other hand, the binary search takes O(log n)
comparisons to search an element in the list but it requires the list
should already be sorted.
Some applications such as direct files, require the search to be of order
O(1) in the best case i.e. the time of search should be independent of the
location of the element or key in the storage area. In fact, in this
situation a hash function is used to obtain a unique address for a given
key or a token. The term hashing means key transformation. The
address for a key is derived by applying an arithmetic function called
hash function on the key itself. The address so generated is used to store
the record of the key as shown in Fig. 10.38.

Fig. 10.38 Direct address generation

In the following section, various hashing functions are discussed.
10.5.1 Types of hashing Functions
Many hashing functions are available in the literature. Most of the
efficient hash functions are complex. For instance, MD5 and SHA-1 are
complex but very efficient hash functions used for cryptographic
applications. However, for searching records or tokens in files or tables,
even simple hash functions give good results. Some of the common
Hashing functions are described below.
1. Midsquare method: The value of the key K is squared and
afterwards a suitable number of digits from the middle of K2 is chosen to

be the address of the record on the disk. Let us assume that 4th and 5th
digit from the right of K2 will be selected as the hash address as shown
below :

It may be noted that the records with K = 5314, 6218, and 9351, would be
stored at addresses 38, 63, and 41 respectively.
2. Division method: In this method the key K is divided by a prime
number where a prime number is a number that is evenly divisible
(without a remainder) only by one or by the number itself. After the
division, the quotient is discarded and the remainder is taken as the
address. Let us consider the examples given below:
(i) Let key K = 189235

Prime number (p) = 41 (say)

Hash address

= K mod p
= 189235 mod 41
= 20 (remainder)

(ii) Let key K = 5314

Prime number (p) = 41

Hash address

= K mod p
= 5314 mod 41
= 25 (remainder)

3. Folding method: In this method, the key is split into pieces and a
suitable arithmetic operation is done on the pieces. The operations can
be add, subtract, divide etc. Let us consider the following examples:
1. Let key K = 189235
Let us split it into two parts 189 and 235
By adding the two parts we get,

189
235
Hash address

424

2. Let key K = 123529164

Splitting the key into pieces and adding them, we get
Let us split it into three parts 123, 529 and 164
By adding the two parts we get,

123
529
164
Hash address

816

It may be observed from the examples given above, that there are
chances that the records with different key values may hash to the same
address. For example, in folding technique, the keys 123529164 and
529164123 will generate the same address, i.e., 816. Such mapping of
keys to the same address is known as a collision and the keys are called
as synonyms.
Now, to manage the collisions, the overflowed keys must be stored in
some other storage space called overflow area. The procedure is
described below:
1. When a record is to be stored, a suitable hashing function is applied on the
key of the record and an address is generated. The storage area is accessed,
and, if it is unused, the record is stored there. If there is already a record
stored, the new record is written in the overflow area.
2. When the record is to be retrieved, the same process is repeated. The record
is checked at the generated address. If it is not the desired one, the system
looks for the record in the overflow area and retrieves the record from there.

Thus, synonyms cause loss of time in searching records, as they are not
at the expected address. It is therefore essential to devise hash
algorithms which generate minimum number of collisions. The
important features required in a hash algorithm are given in the
following section.
10.5.2 requirements for hashing Algorithms
The important features required in a Hash algorithm or functions are :
1. Repeatable: A capability to generate a unique address where a record can
be stored and retrieved afterwards.

2. Even distribution: The records of a file should be evenly distributed

throughout the allocated storage space.
3. Minimum collisions: It should generate unique addresses for different
keys so that number of collisions can be minimized.

Nevertheless, the collisions can be managed by using suitable methods

as discussed in next section:
10.5.3 Overflow Management (Collision handling)
The following overflow management methods can be used to manage the
collisions.
1. Chaining: The overflowed records are stored in a chain of collisions
maintained as linked lists as shown in Figure 10.39

Fig. 10.39 Chaining

Assuming that (K mod 10) was taken as the hash function. Therefore, the
keys with values 0, 10, 20, 30 etc. will map on the same address i.e. 0
and their associated records are stored in a linked list of nodes as shown
in figure. Similarly 19, 9, and 59 will also map on the same address i.e. 9
and their associated records stored in a linked list of nodes.
It may be noted that when a key K is searched in the above arrangement
and if it is not found in the head node then it is searched in the attached
linked list. Since the linked list is a pure sequential data structure, the
search becomes sequential. A poor hash function will generate many
collisions leading to creation of long linked lists. Consequently the search
will become very slow. The remedy to above drawback of chaining is
rehashing.
2. Rehashing: When the address generated by a hash function F1 for a
key Kj collides with another key Ki then another hash function F2 is
applied on the address to obtain a new address A2. The collided Key Kj is
then stored at the new address A2 as shown in Fig. 10. 40

Fig. 10.40 Rehashing

It may be noted that if the space referred to by A2 is also occupied, then
the process of rehashing is again repeated. In fact rehashing is
repeatedly applied on the intermediate address until a free location is
found where the record could be stored.
EXERCISES
1.
2.
3.
4.

When does a binary search tree (BST) become a skewed tree?

What is an AVLTree? What are its advantages over a BST?
Write an explanatory note on AVL rotations.
For the AVLTree shown in Figure 10.41, label each node in the AVL
Tree with the correct balance factor. Use 1 for left higher, 0 for equal
height, and 1 for right higher subtrees.

Fig. 10.41 An AVL tree

5. For the AVLTree shown in Fig. 10.41, draw the final AVLTree that would
result from performing the actions indicated from (a) to (e) listed below:

a.
b.
c.
d.
e.

2.
3.
4.
5.

Insert the key 10.

Insert the key 95.
Insert the key 80, and then insert the key 77.
Insert the key 80, and then insert the key 83.
Insert the key 45.

Note: Always start with the given tree, the questions are not accumulative.
Also, show both the data value and balance factor for each node.
Define set data structure. How are they different from arrays? Give
examples to show the applications of the set data structures.
Write an algorithm that computes the height of a tree.
Write an algorithm that searches a key K in an AVLTree.
What are the various cases of insertion of a key K in an AVLTree?

6. Define the terms: null set, subset, disjoint set, cardinality of a set, and
partition.
7. What are the various methods of representing sets?
8. Write an algorithm that takes two sets S1 and S2 and produces a third set S3
such that S3 = S1 S2.
9. Write an algorithm that takes two sets S1 and S2 and produces a third set S3
such that S3 = S1 S2.
10. Write an algorithm that takes two sets S1 and S2 and produces a third set S3
such that S3 = S1 S2.
11. How are web pages managed by a search engine?
12. Give the steps used for colouring a black and white photograph.
13. Write a short note on spell checker as an application of sets.
14. You are given a set of persons P and their friendship relation R, that is, (a,
b) R if a is a friend ofb. You must find a way to introduce person x to
person y through a chain of friends. Model this problem with a graph and
describe a strategy to solve the problem.
15. The subset-sum problem is defined as follows:
Input: a set of numbers A = {a1, a2, , aN} and a number x;
Output: 1 if there is a subset of numbers in A that add up to x. Write down
the algorithm to solve the subset-sum problem.
16. What is a skip list? What is the necessity of adding extra links?
17. Write an algorithm that searches a key K in a skip list.
18. What is an mwaytree.
19. Define a Btree. What are its properties?
20.Write an algorithm for searching a key K, in a Btree.
21. Write an algorithm for inserting a key K, in a Btree.
22. Write an explanatory note on deletion operation in a Btree.
23. What are the advantages of a Btree.

ppendix A
ASCII Codes (Character Sets)
A.1 ASCII (American Standard Code for Information
Interchange) Character Set

Codes 0031 and 127 are nonprintable control characters.

A.2 EBCDIC (Extended Primary Coded Decimal Interchange
Code) Character Set

Appendix B
Table of Format Specifiers
Format

Specified Meaning

data item is a single character

data item is a decimal integer

data item is a floating point value

data item is a floating point value

data item is a floating point value

h*

data item is a short integer

data item is a decimal, hexadecimal or octal integer

data item is a octal integer

data item is a string followed by a white space

data item is a is an unsigned decimal integer

data item is a hexadecimal integer

[]*

data item is a string which may include while space characters

* Only applicable for data input i.e for scanf() function.

Appendix C
Escape Sequences
Character

Escape Sequence

ASCII Value

bell (alert)

\a

backspace

\b

horizontal table

\t

newline (line feed)

\n

vertical tab

\v

11

form feed

\f

12

carriage return

\r

13

quotation mark ()

34

apostrophe ()

39

question mark (?)

\?

63

backslash (\)

\\

92

null

octal number

\000

(0 represents an octal digit)

Appendix D
Trace of Huffman Algorithm