Patni Computer Systems Ltd.

DBMS/SQL+

Training Workshop

Student Guide

Copyright © 2005 Patni Computer Systems Ltd., Akruti, MIDC Cross Road No. 21,
Andheri (E), Mumbai 400 093. All rights reserved. No part of this publication
can be reproduced in any way, including but not limited to photocopy,
photographic, magnetic, or other record, without the prior agreement and
written permission of Patni Computer Systems.

Patni Computer Systems considers information included in this document to be

Confidential and Proprietary.
DBMS/SQL+

Index
1. INTRODUCTION TO DATABASE .........................................................................3
1.1. Introduction.........................................................................................................3
1.2. Characteristics of DBMS ....................................................................................4
1.3. The DBMS Models or Architectures..................................................................6
1.4. Relational DBMS ..................................................................................................7
2. INTRODUCTION TO SQL .....................................................................................12
2.1. What is SQL ?.....................................................................................................12
2.2. Creation of Database Objects ........................................................................14
2.3. Modification of Existing Database Objects ..................................................17
2.4. Deletion Of Database Objects........................................................................19
2.5. Addition Of Data into Tables..........................................................................20
2.6. Modification of Existing Data in Tables........................................................22
2.7. Deletion of Data from Tables.........................................................................22
2.8. Transaction Control .........................................................................................23
2.9. How to Retrieve Data from Tables ? .............................................................25
2.10. SQL Functions................................................................................................39
3. JOINS AND SUBQUERIES ...................................................................................48
3.1. Joins....................................................................................................................48
3.2. Subqueries .........................................................................................................53
4. ADVANCED SQL.....................................................................................................58
4.1. Index ...................................................................................................................58
4.2. Synonym .............................................................................................................58
4.3. Sequence............................................................................................................59
4.4. View....................................................................................................................59
5. PL/SQL ......................................................................................................................61
5.1. Introduction to PL/SQL ...................................................................................61
5.2. PL/SQL block and its Sections........................................................................61
5.3. Declaration Section..........................................................................................62
5.4. Variable Scope And Visibility .........................................................................65
5.5. Programming Constructs ................................................................................69
6. ERROR HANDLING ...............................................................................................74
6.1. Types of Errors in PL/SQL...............................................................................74
6.2. Declaring Exceptions .......................................................................................74
7. STORED PROCEDURES AND FUNCTIONS .....................................................78
7.1. Procedure ..........................................................................................................78
7.2. Functions ...........................................................................................................82
7.3. Packages ............................................................................................................84

Page 1 of 92
DBMS/SQL+

8. TABLE OF EXAMPLES .........................................................................................88

9. APPENDIX A ............................................................................................................89
10. APPENDIX B ............................................................................................................90

Page 2 of 91
 DBMS/SQL+

1. Introduction to Database
1.1. Introduction

A set of inter-related data is known as database and the software that

manages it is known as database management system or DBMS. Hence DBMS
can be described as "a computer-based record keeping system which consists
of software for processing a collection of interrelated data". A set of
structures and relationships that meet a specific need is called a schema.

A DBMS is a software that manages a collection of interrelated data

elements, stored in a database, that can be accessed in a shared manner by
a collection of applications programs. On a very conceptual level, we can
think of a database as central reservoir of data that can be accessed by
many users

A person known as the Database Administrator or the DBA centrally manages

the database. The DBA initially studies the System and accordingly decides
the types of data to be used, then the structures to be used to hold the data
and the interrelationships between the data structures. He then defines data
to the DBMS. The DBA also ensures the security of the database. The DBA
usually controls access to the data through the user codes and passwords and
by restricting the views or operations that the users can perform on the
database.

DBMS
END USER TOOLS

APPLICATIONS

DATA
END END
USER USER

Figure 1-1 DBMS Structure

Page 3 of 92
 DBMS/SQL+

In today’s environment, there must be one key individual who is responsible for
the overall centralized control of the organization’s data. We refer to the
person as the data base administrator (DBA). It concerns policy-oriented tasks
regarding data strategy and overall data planning. The DBA is responsible for
the actual entities in the business environment that will be represented in the
database. The DBA also works with end users and system analysts to document
the data that will be stored in the database. The DBA initially studies the
System and accordingly decides the types of data to be used, then the
structures to be used to hold the data and the interrelationships between the
data structures. He then defines data to the DBMS. The DBA also ensures the
security of the database. The DBA usually controls access to the data through
the user codes and passwords and by restricting the views or operations that the
users can perform on the database.

1.2. Characteristics of DBMS

Some of the general characteristics of the DBMS have been discussed below:

1.2.1. Control of Data Redundancy

When the same data is stored in a number of files it brings in data
redundancy. In such cases, if the data is changed at one place, the change
has to be duplicated in each of the files.

The main disadvantages of data redundancy are:

• Storage space gets wasted
• Processing time may be wasted as more data is to be handled
• Inconsistencies may creep in

In the traditional file environment, application programmers design their own

data files and establish their own record formats. This makes their programs
dependent on the format of those records, and it tends to create a high degree
of data redundancy. For example, many different programs in an installation
may require access to information about employees. If the records that contain
employee information are designed independently, the same piece of employee
information may exist in different files, possibly in different forms. Through
effective use of database techniques, this redundancy can be better controlled.
Ideally, each piece of information is stored only once in the central pool, and all
application programs that require the information gain access to it via the data
base management system. By controlling redundancy in this manner, it is much
easier for the installation to ensure that multiple copies of the same data
element do not exist in different stages of updating.

Page 4 of 91
DBMS/SQL+

1.2.2. Sharing of Data

DBMS allow many applications to share the data.

1.2.3. Maintenance of Integrity

Integrity of data refers to the correctness, consistency and interrelationship
of data with respect to the application that uses the data. Some of the
aspects of data integrity are:

• Many data items can only take a restricted set of values

• Certain field values are not to be duplicated across records. Such
restrictions, called primary key constraints can be defined to the
DBMS
• Data integrity, which defines the relationships between different
files, is called referential integrity rule, which can also be specified
to the DBMS

1.2.4. Support for Transaction Control and Recovery

Multiple changes to the database can be clubbed together as a single 'logical
transaction'. The DBMS will ensure that the updates take place physically
only when the logical transaction is complete.

1.2.5. Recovery and Restart

When traditional files are used to implement systems, recovery and restart
routines (or programs) must be built into each individual application system.
This makes it difficult to implement sophisticated recovery and restart
procedures. With a database management system, recovery and restart
procedures can be handled automatically on a system wide basis using the
tools provided by DBMS software.

A DBMS provides powerful tools for automatically recovering from system

failures, restoring the database to its original form, and restarting jobs that
were affected by failures.

1.2.6. Data Independence

In conventional file based applications, programs need to know the data
organisation and access technique to be able to access the data. This means
that if you make any change in the way the data is organised you will also
have to take care to make changes to the application programs that apply to
the data. In DBMS, the application programs are transparent to the physical
organization and access techniques.

Page 5 of 91
 DBMS/SQL+

1.2.7. Availability of Productivity Tools

Tools like querying language screen and report painter and other 4GL tools
are available. These tools can be utilized by the end-users to query, print
reports etc. SQL is one such language, which has emerged as standard.

1.2.8. Security
DBMSs provide tools by which the DBA can ensure security of the database.

1.2.9. Hardware Independence

Most DBMSs are available across hardware platforms and operating systems.
Thus the application programs need not be changed or rewritten when the
hardware platform or operating system is changed or upgraded.

1.3. The DBMS Models or Architectures

A database management system generally conforms to one of the three
major database architectures: hierarchical, network, or relational. The
range of data structures supported and the availability of data handling
languages depend on the model of DBMS is based. The models are

1. The Hierarchical Model

2. The Network Model
3. The Relational Model.

1.3.1. The Hierarchical Model

In a hierarchically structured database, data records are connected with
embedded pointers to form an inverted tree structure in which dependent
records can have one and only one parent. A database model in which
records in a file are associated in a one-to-many, or parent-child,
relationship.

1.3.2. The Network Model

A network-structured database takes the form of a mesh structure, in which
dependent records can have more than one parent. Like hierarchical model
the relationship is implemented by means of embedded pointers. A DBMS
model that builds a tight linkage, called a set, between elements of data.

Page 6 of 91
 DBMS/SQL+

1.3.3. The Relational Model

Marks Table
Student Table Course Table
Ccode Scode Marks
Scode Sname Ccode Cname
C1 S1 65
S1 A C1 Physics C2 S1 78
S2 B C2 Chemistry C3 S1 83
C3 Maths
C4 S1 85
C4 Biology
Figure 4 : Relational C3
Model S2 83
C4 S2 85
Figure 1-2 Relational Tables

With relation database, data is represented in the form of tables, and no

embedded pointers are required to implement relationships between
records. Because of lack of linkages relational model is easier to understand
and implement.

1.4. Relational DBMS

1.4.1. Introduction
The relational model presents an orderly, predictable and intuitive approach
to organizing, manipulating and viewing data.

1.4.2. RDBMS Terminology

Relational data consists of relations. A relation (or relational table) is a two
dimensional table with special properties. A relational table consists of a set
of named columns and an arbitrary number of rows. The columns are called
Attributes or Fields and rows are called tuples or records. Each attribute
draws values from a pool of values called domain. Simply put, each attribute
is associated with a domain. Note that it is possible to associate more than
one attribute of a table(s) to a single domain.

Page 7 of 91
DBMS/SQL+

DEPT table
Deptno Dname Loc
10 Accounting New York
20 Research Dallas
30 Sales Chicago 'row' or 'tuple'
40 Operations Boston
á
'column' or 'attribute'
EMPLOYEE table
Empno Empname Job Mgr Deptno
7369 Smith Clerk 7902 20
7499 Allen Salesman 7839 30
7566 Jones Manager 7839 20
7839 King President 10
7902 Ford Analyst 7566 20

1.4.3. Properties of Relational Data Structures

Relational tables have six properties, which must be satisfied for any table
to be classified as relational. These are:

1. Entries of attributes are single valued or atomic

Entry in every row and column position in a table must be single valued.
This means columns do not contain repeating groups. For example:
There is an attribute in an EMPLOYEE table called Empname, which
should not contain values like Smith John as it is not atomic or single
value. Instead you can decompose the attribute Empname to Firstname,
Middlename and the Lastname. Now each of these attributes will
contain single value
2. Entries of attribute are of the same kind or data type.
Entries in a column must be of same kind or data type. A column
supposed to store salary of an employee should not store commission.
For instance, all the values of the attribute MGR in the table EMPLOYEE
should contain only digits not any other characters.
3. No two rows are identical
Each row should be unique, this uniqueness is ensured by the values in a
specific set of columns called the primary key.
4. The order of attributes is unimportant
There is no significance attached to order in which columns are stored in
the table. A user can retrieve columns in any order.

Page 8 of 91
DBMS/SQL+

5. The order of rows is unimportant

There is no significance attached to the order in which rows are stored
in the table. A user can retrieve rows in any order.
6. Every column can be uniquely identified.
Each column is identified by its name and not its position. A column
name should be unique in the table.

1.4.4. Database Schema

Database schema is the organization of information within a database for a
single user or multiple users. It is a store of data that describes the content
and structure of the physical data store. It contains various information like
data types, relationships, access controls, etc. The Schema for a user is a set
of structures and relationships between them that meet a specific need. It is
a repository that has information about the structure and content of the
database. It also contains information about how the data is stored
internally, as well as how it is stored physically on the storage device.

1.4.5. Data Integrity

Integrity refers to the wholeness and soundness of the database. Some of
the most important integrities are discussed below.

1.4.6. Domain Constraints

A domain is the set of all possible data values of some particular type. For
example, the domain of employee numbers is the set of all valid employee
numbers; the domain of employee names is the set of valid letters. Thus,
domains are pools of values, from which the actual values appearing in
attributes (columns) are drawn.

1.4.7. Primary Key and Entity Integrity

It is also possible, though again unusual, for a table to have more than one
unique identifier. In such a case we would say that the table has multiple
candidate keys; we would then choose one of those candidate keys to be the
primary key, and the remainder would then be said to be alternate keys (an
alternate key is thus a candidate key that is not primary key).

Let CK be some subset of the columns of table T. Then CK is a candidate key

for T if and only if satisfies the following property:

Uniqueness: At any give time, no two rows of T have the same value for CK

Page 9 of 91
DBMS/SQL+

Note that the relational model requires every table to have at least one
candidate key, and hence requires every table to have a primary key. This
requirement is equivalent to the requirement that, at any given time, no
two rows in the table are identical.

Primary key is a column or set of columns in a table, which uniquely

identifies a row in a table. No two rows of the table can have the same
values for the primary key. Entity integrity is maintained by ensuring that
none of the columns that make up the primary key can take "NULL"
(unknown) values.

1.4.8. Foreign Key and Referential Integrity

The referential rule simply states that the database must not contain any
unmatched foreign key values (i.e., nonnull foreign key values for which
there does not exist a matching value of the corresponding primary key).

In the EMPLOYEE table the deptno field can be considered as a 'foreign

key'. This means that the EMPLOYEE table cannot contain a value for the
deptno, which does not exist in the table DEPT.

The referential integrity rule specifies that if a foreign key in table A refers
to the primary key in table B, then every value of the foreign key in table A
must be null or must be available in table B.

1.4.9. Delete Restrict Referential Integrity

The delete is “restricted” to the case where there are no matching rows in
the table T2 (it is rejected if any such rows exist).

This means that no primary key value can be deleted if there are some
foreign key values dependent on it. i.e Dept no 10 cannot be deleted
because there are some employees assigned to deptno 10.

1.4.10. Delete Cascade Referential Integrity

This means that if a primary key value is deleted then all foreign key values
dependent on it will be deleted. i.e. if Dept no 10 is deleted then all rows
(employees) having dept no 10 will be deleted.

1.4.11. Update Cascade referential Integrity

This means that if a primary key value is updated then all foreign key values
dependent on it will be updated to the new value of primary key. i.e. if we
change the deptno 10 to 50 then all the employees in deptno 10 will be also
shifted to deptno 50. (Column deptno will be automatically updated)

Page 10 of 91
DBMS/SQL+

1.4.12. Column Constraints

These are the constraints, which specify restrictions on the values a column
can take. These restrictions may be defined with or without other values in
the same row

Page 11 of 91
 DBMS/SQL+

2. Introduction to SQL
2.1. What is SQL ?
SQL or Structured Query Language is the set of commands, which is used
for interacting with Relational Databases. ANSI has standardized the
commands and all relational databases conform to this standard. Most RDBMS
provide extensions to it to make the life of the application programmer
easier.

2.1.1. Components of SQL

SQL can be broadly classified into three components.

1. Data Definition Language (DDL):

This component is used for defining and destroying objects in your
application like Tables, Indexes etc. E.g. DDL statements like CREATE
TABLE
2. Data Manipulation Language (DML):
This is used to modify and retrieve the data according to your needs for
example Inserting, Updating and Querying. E.g. DML statements like
SELECT, INSERT, DELETE and UPDATE.
3. Data Control Language (DCL):
This component is used for defining your security for your application
objects like Tables etc. E.g. DCL statements like GRANT and REVOKE.

A database is a collection of structures with appropriate authorizations and

accesses defined. The tables, indexes are structures in the database are
called objects in the database; their names and those of columns are called
identifiers.

SQL commands are the instructions given for some operation on the database
and the fact that SQL DML like SELECT, UPDATE and DELETE statements
typically operate on entire sets of rows, instead of just one row at a time.
Hence, SQL is called set-level language.

2.1.2. Benefits of SQL

The benefits of SQL are as follows:

Page 12 of 91
DBMS/SQL+

1. Non-Procedural Language:
• Set-level languages such as SQL are sometimes described as
“nonprocedural,” on the grounds that the users specify what, not how
(i.e., they say what data they want without specifying a procedure for
getting it). Perhaps a better way of putting matters is to say that
languages like SQL are at a higher level of abstraction than languages
like COBOL or C.
• It processes sets of records rather than just one record at a time.
• It provides automatic navigation to the data. The process of “navigating”
around the physical database to locate the desired data is performed
automatically by the system, not manually by the user
2. Language For All Users:
• System Administrators
• Database Administrators
• Security Administrators
• Application Programmers
• End-Users

2.1.3. Data types Supported in SQL

Oracle 8 has evolved from an RDBMS to an ORDBMS (Object Relational
Database Management System). The traditional Data Types are called Scalar
Data Types.

SCALAR DATA TYPES

Datatype Description

CHAR(n) To store fixed length string

Maximum length = 2000 bytes
Eg. name CHAR(15)

VARCHAR2(n) To store variable length string

Maximum length = 4000 bytes
Eg. description VARCHAR2(100)
LONG(n) To store variable length string
Maximum length = 2 Gigabytes
Eg. synopsis LONG(5000)

Page 13 of 91
 DBMS/SQL+

NUMBER(p,s) To store numeric data

Range is 1E-129 to 9.99E125
Max Number of significant digits = 38
Eg. salary NUMBER(9,2)
DATE To store DATE. Range from January 1, 4712 BC to
December 31, 9999 AD. Both DATE and TIME are
stored.
Requires 7 bytes.
Eg. hiredate DATE

RAW(n) To store data in binary format such as signature,

photograph.
Maximum size = 255 bytes
LONG RAW(n) Same as RAW. Maximum size = 2 Gigabytes

2.2. Creation of Database Objects

2.2.1. TABLE
Tables are objects, which store the user data. In Oracle 8i, tables can be
permanent or temporary. First we will see permanent tables. In the next
section, temporary tables are explained.

Syntax

CREATE TABLE table_name

(
{col_name.col_datatype [[CONSTRAINT
const_name][col_constraint]]},...
[table_constraint],...
)
[AS query]
A table can have a maximum of 1000 columns. Only one column of type LONG is
allowed per table.
Table_name,col_name,const_na A string upto 30 characters length.
me Can be made up to A-Z,0-9,$,_,#
Must begin with a non-numeric ORACLE data
characters.
Col_datatype One of the previously mentioned types
Col_constraint A restriction on the column
Can be of following types-
PRIMARY KEY, NOT NULL, UNIQUE,
FOREIGN KEY, CHECK, Can be named
Only one PRIMARY KEY allowed per table.

Page 14 of 91
DBMS/SQL+

Only one PRIMARY KEY allowed per table.

Table_constraint A restriction on single or multiple columns. The
types are same as in col_constraint. (NULL
constraint is not allowed here).
AS query Query is an SQL statement using SELECT command.
SELECT command returns the rows from tables. See
SECTION 3.11 for more information on SELECT
command.
Useful if the table being created is based on an
existing table. New table need not have all the
columns of the old table.
Name of columns can be different. All or part of the
data can be copied.

Any object created by a user is accessible to the user and the DBA only. To
make the object accessible to other users, the creator or the DBA must
explicitly give permission to others. SEE Section 3.12 for granting permissions.

Example:
CREATE TABLE emp
(
empno NUMBER(4),
ename VARCHAR2(10) ,
deptno NUMBER(2) ,
job CHAR(9) ,
hiredate DATE );

If a table is created as shown above, then there is no restriction on the data

that can be stored in the table. However, if we wish to put some restriction on
the data, which can be stored in the table, then we must supply some
constraints for the columns.

For example, if we want

1. To make empno as the primary key of the table and
2. To ensure that the ename column does not contain NULL values and
3. The job column to have only UPPERCASE entries and
4. to put the current date as the default date in hiredate column in case data
is not supplied for the column.

The create statement can be rewritten as:

Page 15 of 91
DBMS/SQL+

CREATE TABLE emp

(
empno NUMBER(4) CONSTRAINT P_KEY PRIMARY KEY,
ename VARCHAR2(10) CONSTRAINT ENAME_NOT_NULL NOT NULL,
deptno NUMBER(2),
job CHAR(9) CONSTRAINT JOB_ALL_UPPER
CHECK(job = UPPER(job)),
hiredate DATE DEFAULT SYSDATE
);

Now if we have a table DEPT that has following description,

CREATE TABLE Dept
(
Deptno NUMBER(2) ,
Dname VARCHAR2(14) NOT NULL ,
Loc VARCHAR2(13) NOT NULL
);

In EMP table, for deptno column, if we want to allow only those values that
already exist in deptno column of DEPT table, we must enforce what is known
as REFERENTIAL INTEGRITY.

To enforce referential integrity, declare deptno field of DEPT table as PRIMARY

KEY and deptno field of EMP table as FOREIGN KEY as follows.

CREATE TABLE Dept

(
Deptno NUMBER(2) CONSTRAINT DEPTNO_P_KEY PRIMARY KEY,
Dname VARCHAR2(14) NOT NULL,
Loc VARCHAR2(13) NOT NULL
);

CREATE TABLE Emp

(
Empno NUMBER(4) CONSTRAINT P_KEY PRIMARY KEY,
Ename VARCHAR2(10) CONSTRAINT ENAME_NOT_NULL NOT NULL,
Deptno NUMBER(2),
Job CHAR(9) CONSTRAINT JOB_ALL_UPPER
CHECK (Job =UPPER(Job)),
Hiredate DATE DEFAULT SYSDATE,
CONSTRAINT DEPTNO_F_KEY FOREIGN KEY (Deptno)
REFERENCES Dept(Deptno)
);

Page 16 of 91
 DBMS/SQL+

In above example FOREIGN KEY has been declared as a table constraint. Same
can be given as a column constraint as follows.

CREATE TABLE emp

(
Empno NUMBER(4) CONSTRAINT P_KEY PRIMARY KEY,
Ename VARCHAR2(10) CONSTRAINT ENAME_NOT_NULL NOT NULL,
Deptno NUMBER(2) CONSTRAINT DEPTNO_F_KEY REFERENCES
Dept(Deptno),
Job CHAR(9) CONSTRAINT JOB_ALL_UPPER
CHECK(job=UPPER(job)),
Hiredate DATE DEFAULT SYSDATE
);
The following example shows how to create a new table based on an existing
table

CREATE TABLE Newemp(Emp#,Emp_Name,Hire_Date)

AS
(
SELECT Empno,Ename,Hiredate FROM emp
WHERE Hiredate >'01-jan-82'
);

A new table will be created. It will contain empno, ename, hiredate of all
employees hired after 1st Jan 1982. Constraints on an old table will not be
applicable for a new table

2.3. Modification of Existing Database Objects

2.3.1. TABLE
An existing table can be modified even if contains data. However, it is
recommended that you create a table after putting sufficient thought into it.

ALTER TABLE table_name

[ADD (col_name col_datatype col_constraint ,...)]|
[ADD (table_constraint)]|
[DROP CONSTRAINT constraint_name]|
[MODIFY existing_col_name new_col_datatype new_constraint new_default]
[DROP COLUMN existing_col_name]
[SET UNUSED COLUMN existing_col)name];

table_name must be an existing table.

The rules for adding a column to the table are:

1. Can add any column without a NOT NULL specification.
2. Adding a NOT NULL column is possible in three steps.

Page 17 of 91
DBMS/SQL+

• Add a column without NULL specification

• Filling every row in that column with data
• Modifying the column to be NOT NULL

The rules for modifying the columns are:

1. Can increase a character column's width any time.

2. Can increase the number of digits in a number at any time.
3. Can increase or decrease the number of decimal places in a number column
at any time. Any reduction on precision and scale can be on empty columns
only.
4. You can add only NOT NULL constraint using column constraints. Rest all
constraints have to be specified as table constraints
For adding three more columns to the EMP table.

ALTER TABLE emp

ADD (sal NUMBER(7,2) CONSTRAINT SAL_GRT_0 CHECK(sal >0),
mgr NUMBER(4),
comm NUMBER(9,2));

For adding referential Integrity on mgr column

ALTER TABLE emp
ADD CONSTRAINT EMP_FMGR_KEY FOREIGN KEY (mgr) REFERENCES
EMP (empno);

For modifying the width of sal column.

ALTER TABLE emp

MODIFY (sal NUMBER (8,2)) ;

For dropping the FOREIGN KEY constraint on mgr

ALTER TABLE emp
DROP CONSTRAINT EMP_F_KEY;

Rules for dropping column

Oracle 8i allows you to drop a column from a table. This can be done in two
ways.

Page 18 of 91
 DBMS/SQL+

1. Marking the columns as unused and then later dropping them

ALTER TABLE emp
SET UNUSED COLUMN comm;

ALTER TABLE emp

SET UNUSED (sal, hiredate);

Once all the required columns have been marked as unused, we can use the
following command to remove the columns permanently. Columns once
marked as unused cannot be recovered. Marking the columns as unused does
not release the space occupied by them back to the database. Also until you
drop these columns actually, they continue to count towards the absolute
limit of 1000 columns per table. Also, if you mark a column of data type
LONG as UNUSED, you cannot add another LONG column to the table until
you actually drop the unused LONG column.

ALTER TABLE emp

DROP UNUSED COLUMNS;

The advantage of the above-mentioned steps is that marking the columns is

much faster process than dropping the columns.

You can refer to the data dictionary table USER_UNUSED_COL_TABS to get

information regarding the tables with columns marked as unused.

2. Dropping the columns directly

ALTER TABLE emp

DROP COLUMN sal;

This command should be used with caution.

Before the user issues DROP COLUMN command, the user must have
exclusive control over the table.

When you drop a column

• All indexes defined on any of the target columns are also dropped
• All constraints that reference a target column are removed

2.4. Deletion Of Database Objects

Deletion of objects existing in the database is an easy task. Just say

Page 19 of 91
DBMS/SQL+

DROP Obj_Type obj_name;

For example
DROP TABLE EMP;
A table that is dropped cannot be recovered. When a table is dropped,
dependent objects such as indexes are automatically dropped. Synonyms and
views created on the table remain, but give an error if they are referenced. You
cannot delete a table that is being referenced by another table. To do so use
the following
DROP TABLE table-name CASCADE CONSTRAINTS;
DROP TABLE dept CASCADE CONSTRAINTS;

DROP SYNONYM new_emp;

If new_emp is a synonym for a table, then the table is not affected in any way.
Only the duplicate name is removed.

2.5. Addition Of Data into Tables

INSERT INTO table_name[(col_name1,col_name2,...)]
{VALUES (value1,value2,....) | query};

• If values are specified for all columns in the order specified at creation,
then col_names can be omitted
• Values should match datatype of the respective columns
• Number of values should match the number of column names mentioned
• All columns declared as NOT NULL should be supplied with a value
• Character strings should be enclosed in quotes
• DATE values should be enclosed in quotes
• VALUES will insert one row at a time
• Query will insert all the rows returned by the query
• Table_name can be a table or a view or a synonym name
• If table_name is a view, then following restrictions apply
• View CANNOT have a GROUP BY, CONNECT BY, START WITH, DISTINCT,
UNION, INTERSECT OR MINUS clause or a join
• If view has WITH CHECK OPTION clause, then a row that cannot be returned
by the view will not be inserted

Page 20 of 91
DBMS/SQL+

Inserting a row in emp table giving all values

INSERT INTO emp

VALUES (s1. NEXTVAL,'JONATHAN SEAGULL',25,
'615-3520',007,'01-APR-1993',7500);

• 007 is a dept number which exists in dept table

• s1.NEXTVAL returns the next serial number from sequence s1
Inserting a row in EMP table giving some values.

INSERT INTO EMP (empno, ename, sal)

VALUES (200, 'Archer', 40000);

• This row will be created if all the constraints like NOT NULL are satisfied.
Inserting rows in a table from another table.

INSERT INTO new_emp_table

SELECT * FROM emp WHERE emp.hiredate > '11-jan-82';

This example assumes that new_emp_table exists.

2.5.1. Using Substitution Variables

Instead of using a table name or a column name or a value in the SQL, we can
use a variable.

INSERT INTO dept VALUES (&deptno, ‘&dname’, ‘&loc’);

When the above INSERT statement is executed from the SQL prompt, the user is
prompted with following

Enter value for deptno: 10

Enter value for dname: Sales
Enter value for loc: Mumbai

After entering the above details SQL displays the following

old 1: INSERT INTO dept VALUES(&1,&2,&3)

new 1: INSERT INTO dept VALUES(10,'Sales','Mumbai')
INSERT INTO dept VALUES (10,'Sales','Mumbai')

Deptno, dname and loc variables values will be obtained at runtime and
substituted in the SQL statement before execution of the statement. The

Page 21 of 91
 DBMS/SQL+

ampersand & before the variable name indicates that, what follows is a
variable and not a value.
Since there will be direct substitution, quotes are not required when you give
input to dname and loc as quotes are already present. If the statement was
given as

INSERT INTO dept VALUES (&deptno, &dname, &loc);

then while giving input quotes are required.

2.6. Modification of Existing Data in Tables

UPDATE table_name
SET col_name = value|
col_name = SELECT_statement_returning_single_value|
(col_name,...) = SELECT_statement
[WHERE condition];

Provides automatic navigation to the data. If WHERE is omitted, all the rows
will be updated and if WHERE is specified only those rows that satisfy the
condition will be updated.

UPDATE emp
SET ename = 'ADAM'
WHERE ename = 'ADAMS';
For making salary of SMITH equal to that of employee no 7369.

UPDATE emp
SET SAL = (SELECT SAL FROM emp
WHERE empno = 7369)
WHERE ename = 'SMITH';
All employees other than the PRESIDENT will have the same job, salary and
commission as that of ALLEN.

UPDATE emp
SET (job,sal,comm) = (SELECT job,sal,comm FROM emp
WHERE ename = 'ALLEN')
WHERE NOT job = 'PRESIDENT';

2.7. Deletion of Data from Tables

DELETE [FROM] {table_name | alias }
[WHERE condition];
table_name can be a table or a view.

If WHERE is omitted, all rows from the table are removed and if WHERE is
specified all rows, which satisfy the condition, are removed.

Page 22 of 91
 DBMS/SQL+

FROM can be omitted without affecting the statement.

DELETE FROM emp; -- all rows will be deleted

DELETE FROM emp

WHERE deptno = 007;
Rows of employees attached to dept 007 will be deleted.

DELETE emp
WHERE ename = 'JOHN';

Record of 'JOHN' will be deleted.

2.8. Transaction Control

A transaction is a logical unit of work that contains one or more SQL

statements. A transaction is an atomic unit; the effects of all the SQL
statements in a transaction can be either all committed (applied to the
database) or all rolled back (undone from the database).

Banking is the standard example. A transaction would encompass the

transfer of money from one account to another. The transfer is a transaction
because the debit of money from one account and credit to another must be
performed as a unit--one part cannot fail.
A transaction is not necessarily just one SQL operation; rather, it is a
sequence of several operations, in general, that transforms a consistent
state of the database into another consistent state, without necessarily
preserving consistency at all intermediate points.

A transaction begins with the first executable SQL statement. A transaction

ends when any of the following occurs:

• You issue a COMMIT or ROLLBACK statement

• You execute a DDL statement (such as CREATE, DROP, RENAME, ALTER).
If the current transaction contains any DML statements, Oracle first
commits the transaction, and then executes and commits the DDL
statement as a new, single statement transaction
• A user disconnects from the session. (The current transaction is
committed.)
• A user process terminates abnormally. (The current transaction is rolled
back.)

Page 23 of 91
DBMS/SQL+

2.8.1. COMMIT

Committing a transaction makes permanent the changes resulting from all

successful SQL statements in a transaction. When a transaction is
committed, the following events take place:

• The changes made by the current transaction are made permanent

• The locks acquired by the transaction are released
• The transaction is marked as “complete”

Syntax: COMMIT;

2.8.2. The SAVEPOINT Command

• Savepoints name and mark the current point in the processing of a

transaction
• Savepoints allow to undo parts of a transaction instead of the whole
transaction
• If an error is made in between, it is not necessary to restart the
transaction from the beginning
• The number of active Savepoints per user session is 255

Example:

INSERT into emp(empno,ename,sal)

VALUES(1004,’Amit’,5000);

SAVEPOINT spt1;

UPDATE emp
SET sal=sal+1000
WHERE deptno=10;

ROLLBACK to spt1;

The SAVEPOINT command divides a transaction into smaller sub units or logical
units. Very large, complex transactions might require dozens or hundreds of DML
commands, it may help your application design if you can hold at certain points
and ROLLBACK or COMMIT just one part of the total transaction, an example of
this is a complex user information input such as entering mathematical
modeling data. The SAVEPOINT command has a parameter, which is the name of
the savepoint. This name can be used to rollback all updates from the time

Page 24 of 91
 DBMS/SQL+

when you issued the named savepoint. You can reuse the savepoint name, each
time you redeclare an existing savepoint the old savepoint of that name is
deleted.

2.8.3. The ROLLBACK Command

Changes made to the database without COMMIT may be abandoned using the
ROLLBACK statement. When a transaction is rolledback, it is as if the
transaction never occurred. When a transaction is rolled back, the following
events take place

• Any changes made to the database is undone

• The locks acquired by the transaction are released
• The transaction is ended

Syntax: ROLLBACK [TO SAVEPOINT <savepoint name>];

DELETE FROM Dept WHERE deptno = 10

SAVEPOINT delete_finished

INSERT INTO Dept VALUES(50,’Mkt’,’Michigan’);

ROLLBACK TO delete_finished

ROLLBACK

The ROLLBACK command removes pending database updates. Issuing a

ROLLBACK without specifying a savepoint name removes all pending updates
since the last transaction start event whereas issuing a ROLLBACK with a
savepoint name parameter removes all pending updates since the time the
specified savepoint was created. The first ROLLBACK above removes all
updates associated with the INSERT command. The second ROLLBACK
removes all updates associated with the DELETE command.

2.9. How to Retrieve Data from Tables ?

SELECT command is used to retrieve rows from table or views. Rows can be
retrieved from a single table or multiple tables.

The syntax of the SELECT statement is

SELECT [ALL | DISTINCT] { * | col_name,...}

Page 25 of 91
DBMS/SQL+

FROM table_name alias,...

[ WHERE expr1 ]
[ CONNECT BY expr2 [ START WITH expr3 ] ]
[ GROUP BY expr4 ] [ HAVING expr5 ]
[ UNION | INTERSECT | MINUS SELECT ... ]
[ ORDER BY expr | ASC | DESC ];
Each clause is evaluated on the result set of a previous clause. The final result
of the query will be always a result table. Only FROM clause is essential. The
WHERE, GROUP BY, HAVING, ORDER BY, UNION are optional.
All the examples that follow are based on EMP and DEPT tables that are already
available to you. For more information about these tables, refer Appendix B.

To select all rows from table dept.

SELECT deptno,dname,loc FROM dept;

DEPT NO DNAME LOC

------- ------------ --------
10 ACCOUNTING NEW YORK
20 RESEARCH DALLAS
30 SALES CHICAGO
40 OPERATIONS BOSTON

• It is NOT necessary to retrieve all the COLUMNS of the table. If LOC is not
required, then the following query can be used

SELECT deptno,dname FROM dept;

• It is NOT necessary to list the column names in the same order as done at
the time of table creation. If dept name is to be displayed followed by dept
number, then use

SELECT dname,deptno FROM dept;

• If all columns are to be retrieved, then instead of listing the column names,
use. The star or asterisk is shorthand for a list of all column names in the
table(s) named in the FROM clause, in the left-to-right order in which those
columns appear in the relevant tables(s)

SELECT * FROM dept;

• When rows are retrieved and displayed, the column heading is same as the
column name. To change the heading from LOC to LOCATION, use

SELECT dname,loc "LOCATION" FROM dept;

• To list deptno from EMP table

Page 26 of 91
DBMS/SQL+

SELECT deptno FROM emp ;

DEPTNO
------------
20
30
30
20
30
30
10
20
10
…
14 rows selected

DISTINCT Parameter
In the above output, some of the values have been repeated. By default, all
values are retrieved. If you wish to remove duplicate values, then say

SELECT DISTINCT deptno FROM emp;

Note: It just so happens in this particular example that each row contains a
single value; the effect of the DISTINCT specification is therefore to eliminate
duplicate values. In general, however, DISTINCT means “eliminate duplicate
rows.

2.9.1. WHERE Clause

WHERE clause is used to do selective retrieval of rows. It follows FROM
clause and specifies the search condition. The result of the WHERE clause is
the row or rows retrieved from the tables which meet the search condition.

The syntax is of the form

WHERE <search condition>

COMPARISON Predicate
The comparison predicates specifies comparison of two values. It is of the form
< Expression> < operator > < Expression>
< Expression> <operator> <subquery> --- Will be discussed Later
The operators used are

Page 27 of 91
DBMS/SQL+

= Equal to
<> not equal to
< Less than
> Greater than
<= Less than or equal to
>= Greater than or equal
to

To list all employees working in department number 10 from EMP table

SELECT ename "NAME”, deptno FROM emp

WHERE deptno = 10;
NAME DEPTNO
----------- ------------
CLARK 10
KING 10
MILLER 10
To list all employees whose sal is greater than or equal to 2000

SELECT ename,sal from emp where sal >=2000;

ENAME SAL
---------- ----------
CLARK 2450
BLAKE 2850
JONES 2975
SCOTT 3000
FORD 3000
KING 5000

BETWEEN Predicate – Retrieval Using BETWEEN

<expression> BETWEEN < expression> AND <expression>
A BETWEEN predicate X BETWEEN Y AND Z is really same as
X >= Y AND X <= Z

The BETWEEN condition is really just shorthand for a condition involving two
individual comparisons “ANDed” together.

Page 28 of 91
DBMS/SQL+

• To list names and sal of all employees whose sal is between 2000 and 3000
including 2000 and 3000

SELECT ename,sal FROM emp WHERE sal BETWEEN 2000 AND 3000;

ENAME SAL
---------- ----------
CLARK 2450
BLAKE 2850
JONES 2975
SCOTT 3000
FORD 3000

To list all employees hired during the year '82.

• SELECT ename, hiredate, sal FROM emp

WHERE hiredate BETWEEN '01-jan-82' AND '31-dec-82';

ENAME HIREDATE
----------- ----------------
SCOTT 09-DEC-82
MILLER 23-JAN-82

IN Predicate – Retrieval Using IN

<Expression> IN <LIST>
<Expression> IN <SUBQUERY>

IN, like BETWEEN, is really just shorthand. An IN condition is logically equivalent

to a condition involving a sequence of individual comparisons all “ORed”
together.
The data types should match. On right hand side a subquery can be used.

• To get all the employees from deptno 10 and 20

SELECT empno, ename, sal, deptno FROM emp

WHERE deptno in (10, 20);
EMPNO ENAME SAL DEPTNO
---------- ---------- ---------- ------------
7369 SMITH 800 20
7566 JONES 2975 20
7782 CLARK 2450 10
7788 SCOTT 3000 20
7839 KING 5000 10
7876 ADAMS 1100 20
7902 FORD 3000 20

Page 29 of 91
DBMS/SQL+

7934 MILLER 1300 10

LIKE Predicate –Retrieval Using LIKE

<COLUMN > LIKE < PATTERN>
The pattern contains a search string along with other special characters % and
_. The % character (percent) represents any sequence of n characters (where n
may be zero) where as the character _ (underscore) represents any single
character. A pattern %XYZ% means search is to be made for string XYZ in any
position. A pattern '_XYZ%' means search is to be made for a string XYZ in
position 2 through 4 with any single character in position 1 and 0 or more
characters starting from position 5.

To search for characters % and _ in the string itself we have to use an escape
character. For example to search for string NOT_APP in column status we have
to use the form
Status like 'NOT\_APP' ESCAPE \' ' The use of \ as escape character is purely
arbitrary.

• To list all employees whose name begins with 'J'

SELECT ename FROM emp WHERE ename LIKE 'J%';

ENAME
------------
JONES
JAMES

• To list employees whose name begins with 'J' and has 'N' as the 3rd
CHARacter, say
SELECT ename FROM emp WHERE ename LIKE 'J_N%';
ENAME
----------
JONES

NULL Predicate –Retrieval Involving NULL

The NULL SQL keyword is used to represent either a missing value or a value
that is not applicable in a relational table. A null can be assigned but it cannot
be equated with anything.
The NULL predicate specifies a test for null values. The form for NULL
predicate is:
< COLUMN SPECIFICATION > IS NULL.
< COLUMN SPECIFICATION > IS NOT NULL.
< COLUMN SPECIFICATION > IS NULL
returns true only when column has null values.

Page 30 of 91
DBMS/SQL+

<COLUMN> = NULL CANNOT BE USED TO COMPARE NULL VALUES.

• To list all employees not entitled for commission.

SELECT ename, sal, comm FROM emp WHERE comm IS NULL;

ENAME SAL COMM

---------- --------- -----------
SMITH 800
JONES 2975
BLAKE 2850
CLARK 2450
SCOTT 3000
KING 5000
ADAMS 1100
JAMES 950
FORD 3000
MILLER 1300

• To list all employees who receive commission

SELECT ename,sal,comm FROM emp WHERE comm IS NOT NULL;

ENAME SAL COMM
---------- ---------- ----------
ALLEN 1600 300
WARD 1250 500
MARTIN 1250 1400
TURNER 1500 0

Even though TURNER has not received any comm, the row has been retrieved
because it contains a 0 and not a NULL. NULL is a special representation without
any value.

• To list the BIG BOSS of the company, except for BIG BOSS, all employees
have a value in the MGR column.

SELECT ename, job,salary FROM emp WHERE mgr IS NULL;

ENAME JOB SAL

----------- -------- -------

KING PRESIDENT 5000

Page 31 of 91
DBMS/SQL+

Few More Predicates

Predicates EXISTS, ANY, ALL will be discussed in SUBQUERY sections.

Combining Predicates Using Logical Operators

Above predicates can be combined using logical operators AND, OR, NOT. The
evaluation proceeds from left to right and order of evaluation is enclosed in
parenthesis

• To list employees working in dept NUMBER 10 or 20, say

SELECT ename "NAME",deptno

FROM emp
WHERE deptno = 10 OR deptno = 20;

• To list employees hired after 01/09/81 and working in dept NUMBER 10, use

SELECT ename ,hiredate,deptno FROM emp

WHERE deptno = 10 AND hiredate >'01-SEP-81';

• To list employees working in depts other than 10, use

SELECT ename FROM emp

WHERE NOT deptno = 10;
NOT is a negation operator.

Aggregate Functions
Although quite powerful in many ways, the SELECT statement as so far
described is still inadequate for many practical problems. For example, even a
query as simple as “How many employees are there?’ cannot be expressed using
the constructs introduced up till now. SQL therefore provides a number of
special aggregate (or column) functions to enhance its basic retrieval power.
Some of the aggregate functions available are COUNT, SUM, AVG, MAX and MIN.
Apart from the special case of “COUNT(*)”, each of these functions operates on
the collection of scalar values in one column of some table.

SQL provides a set of built-in functions for producing a single value for an entire
group. These functions are called Set functions or Aggregate functions.
These functions can work on a normal result table or a grouped result table. If
the result is not grouped then the aggregate will be taken for the whole result
table.

Page 32 of 91
DBMS/SQL+

1. COUNT(*)

COUNT returns the number of rows in a table.

To find the total number of employees

SELECT COUNT(*) FROM emp;

COUNT(*)
----------------
14
It is possible to restrict the rows over which COUNT operates.
To find the total number of CLERKS, use

SELECT COUNT(*) FROM emp WHERE job = 'CLERK' ;

To find the total number of CLERKS hired after '13-jan-81', say

SELECT COUNT(*) FROM emp

WHERE job = 'CLERK' AND hiredate >'13-jan-81' ;
When an aggregate function is used in a SELECT statement, column names
cannot be used in SELECT unless GROUP BY clause is used.

2. SUM(COL_NAME | EX PRESSION)

SUM returns the total of values present in a particular column or a number of

columns linked together in the expression. All the columns, which form the
argument to SUM, must be numeric only.

To find the sum paid as salary to all employees every month

SELECT SUM(sal) FROM emp;

SUM(SAL)
---------------
29025

To find the yearly compensation paid to all SALESMEN use

SELECT SUM(12*sal) FROM emp

WHERE job = 'SALESMAN';

3. AVG (COL_NAME | EXPRESSION)

Page 33 of 91
DBMS/SQL+

AVG is similar to SUM. AVG returns the average of the values in the column. The
restrictions, which apply on SUM also, apply on AVG.
To find the average salary of all employees
SELECT AVG(sal) FROM emp;

AVG (SAL)
---------------
2073.21429

To find the average yearly compensation paid to SALESMEN, use

SELECT AVG(12*sal) FROM emp

WHERE job = 'SALESMAN';
4. MIN (COL_NAME | EXPRESSION)

MIN returns the smallest value in the column. MIN accepts columns, which are
NON-NUMERIC too.
To find the minimum salary paid to any employee
SELECT MIN(sal) FROM emp;

MIN(SAL)
--------------
800
To list the employee who heads the list alphabetically, use
SELECT MIN(ename) FROM emp;

5. MAX (COL_NAME | EXPRESSION)

MAX is the reverse of MIN. MAX returns the maximum value from among the list
of values in the column.

To find the maximum salary paid to any employee.

SELECT MAX(sal) FROM emp ;

MAX(SAL)
---------------
5000

2.9.2. GROUP BY and HAVING

The GROUP BY operator causes the table represented by the FROM clause to
be rearranged into partitions or groups, such that within one group all rows
have the same value for the GROUP BY column. Note that the table is not
physically arranged in the database.

Page 34 of 91
DBMS/SQL+

All the SELECT statements we have used until now have acted on data as if
the data is in a single group. But the rows of data in some of the tables can
be thought of as being part of different groups.

For example, the EMP table contains employees who are CLERKS, SALESMEN
etc. If we wish to find the minimum salary of each group of employees, then
none of the clauses that we have seen until now are of any use.

The GROUP BY are used to group the result table derived from earlier FROM
and WHERE clause. HAVING is used to apply search condition on these
groups.

When a GROUP BY clause is used, each row of the resulting table will
represent a group having same values in the column(s) used for grouping.
HAVING clause then acts on the resulting grouped table to remove the row
which does not satisfy the criteria in the HAVING search condition.
The GROUP BY clause is of the form
GROUP BY < column list >
The columns specified must be selected in the query. If a table is grouped,
the select list should have columns and expressions, which are single-valued
for a group. These are

• Columns on which grouping is done

• Constants
• Aggregate functions on the other columns on which no grouping is done
A HAVING clause is of the form

HAVING <search condition>

The search condition applies to each group. It can be formed using various
predicates like (between, in, like, null, comparison) etc. Combined with
Boolean (AND, OR, NOT) operators. Since the search condition is for grouped
table. The predicates should be

• On a column by which grouping is done.

• a set function ( Aggregate function ) on other columns.
The aggregate functions can be used in HAVING clause. But they cannot be
used in the WHERE clause.

Page 35 of 91
DBMS/SQL+

To Find Out Average and Minimum, Maximum salary of each department.

Select deptno,AVG(sal),MIN(sal),MAX(sal) FROM emp

group by deptno;
DEPTNO AVG(SAL) MIN(SAL) MAX(SAL)
--------- -------------- -------------- --------------
10 2916.66667 1300 5000
20 2175 800 3000
30 1566.66667 950 2850
Explanation of the above query: Here the table emp is grouped so that one
group contains all the rows for dept d1 (say), another contain all rows for dept
d2 (say), and so on. The SELECT clause is then applied to each group of the
partitioned table (rather than to each row of the original table). Each
expression in the SELECT clause must be single-valued per group; e.g., it can be
(one of) the column(s) named in the GROUP BY clause, or a literal, or an
aggregate function such as SUM that operates on all values of a given column
within a group and reduces those values to a single value.

To Find Out Average and Maximum, Minimum salary of departments where

average salary is greater than 2000

SELECT deptno,AVG(sal),MIN(sal),MAX(sal) FROM emp

GROUP BY deptno HAVING AVG(sal) > 2000;
DEPTNO AVG(SAL) MIN(SAL) MAX(SAL)
------------ -------------- --------------- ---------------
10 2916.66667 1300 5000
20 2175 800 3000

To list the minimum salary of various categories of employees.

SELECT job,MIN(sal) FROM emp

GROUP BY job;
JOB MIN(SAL)
------- ----------------
ANALYST 3000
CLERK 800
MANAGER 2450
PRESIDENT 5000
SALESMAN 1250
Only the column names, which have been used in GROUP BY clause and
aggregate columns, can be used in SELECT clause. Grouping can be done on
multiple columns also. To find the minimum salary of various categories of
employees in various departments, use

Page 36 of 91
DBMS/SQL+

SELECT deptno,job,MIN(sal) FROM emp

GROUP BY deptno,job;
The output is sorted on the basis of deptno and within each deptno, the output
is sorted on the order of job. Reverse the order of deptno and job and see the
changes to the output.
To list the minimum salary of various categories of employees, department
wise, such that minimum salary is greater than 1500

SELECT job,deptno,MIN(sal) FROM emp

GROUP BY job,deptno HAVING MIN(sal) > 1500;

JOB DEPTNO MIN(SAL)

-------- ------------ --------------
ANALYST 20 3000
MANAGER 10 2450
MANAGER 20 2975
MANAGER 30 2850
PRESIDENT 10 5000

To find the minimum salary of MANAGER's in various departments

SELECT job, deptno, MIN(sal) FROM emp WHERE job = 'MANAGER'

GROUP BY job,deptno;
JOB DEPTNO MIN(SAL)
--------------- ------------ --------------
MANAGER 10 2450
MANAGER 20 2975
MANAGER 30 2850
It is possible to combine a WHERE clause with a HAVING clause. In above
example if you are interested in retrieving data where minimum salary is
greater than 2500, use
SELECT job, deptno, MIN(sal) FROM emp WHERE job = 'MANAGER'
GROUP BY job,deptno HAVING MIN(sal) >2500 ;

2.9.3. ORDER BY Clause

A query with its various clauses (FROM, WHERE, GROUP BY, HAVING)
determines the rows to be selected and the columns. The order of rows is
not fixed unless an ORDER BY clause is given
An ORDER BY clause is of the form
ORDER BY < Sort list> ASC/DESC

Page 37 of 91
DBMS/SQL+

The columns to be used for ordering are specified by using the column names or
by specifying the serial number of the column in the select list. The sort is done
on the column in Ascending or Descending order. The default is Ascending.
Note: GROUP BY does not imply ORDER BY.
To list all employees in the ascending order by name
SELECT ename, sal from emp order by ename;

ENAME SAL
---------- ----------
ADAMS 1100
ALLEN 1600
BLAKE 2850
CLARK 2450
FORD 3000
JAMES 950
JONES 2975
KING 5000
MARTIN 1250
MILLER 1300
SCOTT 3000
SMITH 800
TURNER 1500
WARD 1250

To select all employees sorted departmentwise in ascending order and within

department, salarywise in descending order.

SELECT deptno,ename,sal FROM emp

order by deptno, sal desc;

DEPTNO ENAME SAL

----- ----------- ----------
10 KING 5000
10 CLARK 2450
10 MILLER 1300
20 SCOTT 3000
20 FORD 3000
20 JONES 2975
20 ADAMS 1100
20 SMITH 800
30 BLAKE 2850
30 ALLEN 1600
30 TURNER 1500
30 WARD 1250
30 MARTIN 1250
30 JAMES 950

To select all employees sorted dept wise in ascending order and within dept
salary wise in descending order for deptno 10 and 20.

Page 38 of 91
 DBMS/SQL+

SELECT deptno,ename,sal FROM emp

WHERE deptno=10 or deptno = 20
ORDER BY deptno,sal DESC;

DEPTNO ENAME SAL

------ ------ ----------
10 KING 5000
10 CLARK 2450
10 MILLER 1300
20 SCOTT 3000
20 FORD 3000
20 JONES 2975
20 ADAMS 1100
20 SMITH 800

To select all employees along with their annual salary sorted on the basis of the
annual salary.
SELECT ename, sal * 12 "Annual Salary" FROM emp
ORDER BY 2 DESC;

Ename Annual Salary

----------- -------------------
KING 60000
SCOTT 36000
FORD 36000
JONES 35700
BLAKE 34200
CLARK 29400
ALLEN 19200
TURNER 18000
MILLER 15600
WARD 15000
MARTIN 15000
ADAMS 13200
JAMES 11400
SMITH 9600

2.10. SQL Functions

We have seen some of the aggregate functions provided by ORACLE. ORACLE
also provides a rich set of functions, which act on single rows.

If you call a SQL function with a null argument, the SQL function
automatically returns null.

Different functions work on different type of data. Following is a list of most

commonly used functions from each category.

Page 39 of 91
DBMS/SQL+

2.10.1. String functions

1. UPPER(string)

• Converts all characters in string to uppercase.

SELECT UPPER(ename),empno FROM emp WHERE job = 'CLERK';
UPPER(ENAME) EMPNO
------------ -----------
SMITH 7902
ADAMS 7788
JAMES 7698
MILLER 7782

2. LOWER(string)

• Converts all characters in string to lowercase.

SELECT LOWER(ename),empno FROM emp WHERE job = 'CLERK';
LOWER(ENAME) EMPNO
------------- ----------
smith 7902
adams 7788
james 7698
miller 7782

3. INITCAP(string)

Converts the first character of each word in string to uppercase and the rest of
the characters to lowercase.
SELECT INITCAP(ename),empno FROM emp
WHERE job = 'CLERK';
INITCAP(ENAME) EMPNO
----------------- -----------
Smith 7902
Adams 7788
James 7698
Miller 7782

4. LPAD(string1,n,string2)

Adds string2 before string1 as many times as required to make the string1 length
equal to n chars.
To right align the names of employees

SELECT empno,LPAD(ename,8,' ') FROM emp WHERE job='CLERK';

Page 40 of 91
DBMS/SQL+

LPAD(ENAME,8,' ') EMPNO

-------------------------- -----------
SMITH 7902
ADAMS 7788
JAMES 7698
MILLER 7782

5. LTRIM(string,CHAR set)

Removes chars from beginning of string as long as the character matches one of
the chars in the CHAR set.
SELECT ename,LTRIM(ename,'MALICE') FROM emp;

ENAME LTRIM(ENAME,
----------- -----------------------
SMITH SMITH
ALLEN N
WARD WARD
JONES JONES
MARTIN RTIN
BLAKE BLAKE
CLARK RK
SCOTT SCOTT
KING KING
TURNER TURNER
ADAMS DAMS
JAMES JAMES
FORD FORD
MILLER R
14 rows selected.

6. RPAD(string1,n,string2)

Similar to LPAD. Adds to the right end.

7. RTRIM(string,CHAR set)

Similar to LTRIM. Removes at the right end.

8. SUBSTR(CHAR,m,n)

This function returns the string from the m th character of the string to the n th
character.

SELECT SUBSTR(‘abcdefg’,3,2) “SUBSTRING” from dual;

SUBSTRING
------------------
cd

Page 41 of 91
DBMS/SQL+

NOTE: The table named DUAL is a small table in the data dictionary that Oracle
and user-written programs can reference to guarantee a known result. This
table has one column called DUMMY and one row containing the value X.

2.10.2. Date and Time Functions

1. SYSDATE

Returns the current DATE and TIME.

SELECT SYSDATE FROM DUAL;

2. TRUNC (DATE1)

Truncates the time part from the DATE. This is required when we do DATE
calculations.

SELECT ename, TRUNC(sysdate) - TRUNC(HIREDATE) FROM emp;

3. MONTHS_BETWEEN (DATE1,DATE2)

Returns number of months between the two DATEs.

SELECT empno,MONTHS_BETWEEN(TRUNC(sysdate),hiredate)
FROM emp;

EMPNO MONTHS_BETWEEN(TRUNC(SYSDATE),HIREDATE)
--------- -------------------------------------------------------------
7369 10.36
7499 1.562
7521 0.721
. .
14 rows selected

4. ADD_MONTHS(DATE1,int1)

Returns a DATE, int1 times added.

SELECT ename,ADD_MONTHS(hiredate,2) FROM emp;
ENAME ADD_MONTHS(HIREDATE,2)
----------- -----------------------------------------
SMITH 17-FEB-81
ALLEN 20-APR-81
. .
. .
14 rows selected

Page 42 of 91
DBMS/SQL+

int1 can also be a negative integer.

5. TO_CHAR(DATE1,format)

Converts the DATE given to the format specified

Give below is a list of some of the formats allowed :

Format Meaning
YYYY Four Digit Year
YY Last two digits of the year
MM Month (01-12;JAN=01...)
MONTH Name of the month stored as a length of nine characters
MON Name of the month in three letter format
DD Day of the month (01-31)
D Day of the week
DAY Name of the day stored as a length of nine characters
DY Name of the day in three letter format
FM This prefix can be added to suppress blankpadding
HH Hour of the day
HH24 Hour in the 24 hr format
MI Minutes of the hour
SS Seconds of the minute
TH The suffix used with the day

SELECT TO_CHAR(TRUNC(sysdate), ‘ddth fmMonth yy’) 'DATE'

FROM dual;
DATE
---------
01st January 95

SELECT TO_CHAR(TRUNC(sysdate),'fmMonth') "DATE" FROM dual;

DATE
---------
July

6. ROUND(n,m)

Page 43 of 91
DBMS/SQL+

Returns n rounded to m places.

If m is omitted, to zero places. m can be negative to round off digits left of the
decimal point.
SELECT ROUND(17.175,1) "Number" FROM dual;

Number
--------
17.2

SELECT ROUND(17.175,-1) "Number" FROM dual;

Number
-----------
20

7. LASTDAY(d)

Returns the date of the last day of the month that contains d. You might use
this function to determine how many days are left in the current month.
SELECT SYSDATE,
LAST_DAY(SYSDATE) "Last",
LAST_DAY(SYSDATE) - SYSDATE "Days Left"
FROM DUAL;

SYSDATE Last Days Left

--------- -------------- ----------
23-OCT-97 31-OCT-97 8

8. NEXT_DAY(d, char)

Returns the date of the first weekday named by char that is later than the date
d.

SELECT NEXT_DAY('15-MAR-98','TUESDAY') "NEXT DAY"

FROM DUAL;
NEXT DAY
---------
16-MAR-98

Numeric Functions
9. TRUNC(n,m)

Returns n truncated to m places. If m is omitted, n is truncated to zero places. m

can be negative to truncate left of the decimal point.

Page 44 of 91
DBMS/SQL+

SELECT TRUNC(15.81,1) "Number" FROM dual;

Number
----------
15.8
SELECT TRUNC(15.81,-1) "Number" FROM dual;

Number
-----------
10

10. ROUND(N,M)

Returns n rounded to m places right of the decimal point.

SELECT ROUND(15.193,1) "Round" FROM DUAL;

Round
----------
15.2
11. CEIL(n)

Returns smallest integer greater than or equal to n.

SELECT CEIL(15.7) "Ceiling" FROM DUAL;

Ceiling
----------
16

12. FLOOR(n)

Returns largest integer equal to or less than n.

SELECT FLOOR(15.7) "Floor" FROM DUAL;

Floor
----------
15
13. ABS(n)

Returns the absolute value of n.

SELECT ABS(-15) "Absolute" FROM DUAL;

Absolute
----------
15

There are some more numeric functions available, Please refer to the online
documentation for more help.
Page 45 of 91
DBMS/SQL+

Other SQL Functions are

LENGTH: Returns the length of char in characters

a) LENGTH (string)

Select LENGTH (‘candide’) “ length in characters”

From dual;

7 Length in characters

b) NVL(exprl,n)

NVL substitutes n if value of exprl is NULL

SELECT ename,NVL(COMM,0) “COMMISSION” FROM emp

WHERE deptno=30;

ENAME COMMISSION
ALLEN 300
WARD 500
MARTIN 1400
BLAKE 0
TURNER 0
JAMES 0

2.10.3. DECODE:
A Special function used to evaluate expression. This works like the IF-THEN-
ELSE conditional loop.

Decode(col_name,condition1,value1,condition2,value2…..,default value)

Select Ename,job,DECODE(job,’SALESMAN’,’ELIGIBLE’,’PRESIDENT’,’HEAD
PERSON’,’NOT ELIGIBLE’) “COMM ELIGIBLITY” FROM EMP;

ENAME JOB COMM ELIGIBILITY

SMITH CLERK NOT ELIGIBLE
SMITH SALESMAN ELIGIBLE
WARD SALESMAN ELIGIBLE
JONES MANAGER NOT ELIGIBLE
MARTIN SALESMAN ELIGIBLE
BLAKE MANAGER NOT ELIGIBLE
CLARK MANAGER NOT ELIGIBLE
SCOTT ANALYST NOT ELIGIBLE
KING PRESIDENT HEAD PERSON

Page 46 of 91
DBMS/SQL+

TURNER SALESMAN ELIGIBLE

ADAMS CLERK NOT ELIGIBLE
JAMES CLERK NOT ELIGIBLE
FORD ANALYST NOT ELIGIBLE
MILLER CLERK NOT ELIGIBLE

POWER: Returns m raised to nth power ------ power(m,n)

POWER(m,n)
SELECT POWER (3,2) “Raised” FROM DUAL;

Raised
9
REPLACE: Returns char with every occurrence of search string replaced with
replacement string. Refer the online documentation for more detail.

a) REPLACE (String, if , then)

SELECT ename,replace (ename, ‘I’ , ‘x’) “changed name” FROM Emp;

ENAME Changed Name

KING KxNG
BLAKE BLAKE
CLARK CLARK
JONES JONES
MARTIN MARTxN
ALLEN ALLEN
TURNER TURNER
JAMES JAMES
WARD WARD
FORD FORD
SMITH SMxTH
SCOTT SCOTT
ADAMS ADAMS
MILLER MxLLER

b) INSTR(string, pattern[ , start [,occurrence ] ] )

Returns the location of a character IN a STRING
SELECT INSTR (‘CORPORATEFLOOR’,’R’,3,2) “Instring” FROM DUAL;

Instring
6

Page 47 of 91
 DBMS/SQL+

3. Joins and Subqueries

3.1. Joins
The ability to “join” two or more tables is one of the most powerful features
of relational systems. In fact, it is the availability of the join operation,
almost more than anything else that distinguishes relational from non-
relational systems. So what is a join? Loosely speaking, it is a query in which
data is retrieved from more than one table.

JOINS make it possible to select data from more than one table by means of
a single statement.

The joining of tables is done in SQL by specifying the tables to be joined in

the FROM clause of the SELECT statement. When you join two tables a
Cartesian product is formed. The conditions for selecting rows from the
product are determined by the predicates in the WHERE clause. All the
subsequent WHERE, GROUP BY, HAVING, ORDER BY clauses work on this
product.

If the same table is used more than once in a FROM clause then to resolve
conflicts and ambiguities aliases are used. They are also called co-relation
names or range variables. Once aliases are used, table names cannot be
used in further clauses.

Assume two tables

TABLE F1
COL1 COL2
----- -----
A 1
B 2
C 3

Table F2
COL1 COL2
----- -----
X 100
Y 200

Page 48 of 91
DBMS/SQL+

The statement SELECT * FROM F1,F2; results in

COL1 COL2 COL1 COL2

--------- ---------- --------- ---------
A 1 X 100
B 2 X 100
C 3 X 100
A 1 Y 200
B 2 Y 200
C 3 Y 200
6 rows selected.
Statement
SELECT * FROM F1 First_T, F1 Second_T;
results in
COL1 COL2 COL1 COL2
--------- ------- ----- ----------
A 1 A 1
B 2 A 1
C 3 A 1
A 1 B 2
B 2 B 2
C 3 B 2
A 1 C 3
B 2 C 3
C 3 C 3
9 rows selected.

3.1.1. Equi-join(Inner Join)

Typically the tables are joined to get meaningful data. Suppose we want to
get the department name of all employees along with their names. Now
since the employee name is in the employee table and dept name is present
in dept table we have to take the join of two tables on the basis of the
common column between these two tables i.e deptno column.

SQL> SELECT ename, dname, sal FROM emp,dept

WHERE emp.deptno = dept.deptno;

ENAME DNAME SAL

----------- -------------- ----------
SMITH RESEARCH 800
ALLEN SALES 1600
WARD SALES 1250
JONES RESEARCH 2975
MARTIN SALES 1250
BLAKE SALES 2850
CLARK ACCOUNTING 2450

Page 49 of 91
DBMS/SQL+

SCOTT RESEARCH 3000

KING ACCOUNTING 5000
TURNER SALES 1500
ADAMS RESEARCH 1100
JAMES SALES 950
FORD RESEARCH 3000
MILLER ACCOUNTING 1300
14 rows selected.

The join is based on the equality of column values in the two tables
(emp.deptno= dept.deptno) and therefore is called an equi-join.

The equi-join is also used to provide summary information. Suppose we wanted

to know the details of all departments along with number of employees working
in it. Since the number of employees can be found out only through the emp
table we have to take a join. Also since we want number of employees per
dept, we have to group the product of the join.

SQL> SELECT dept.deptno, dept.dname, dept.loc, count(*) "NO of

employees" FROM dept,emp
WHERE emp.deptno=dept.deptno
GROUP BY dept.deptno,dept.dname,dept.loc;

DEPTNO DNAME LOC NO of Employees

----------- ------------ ---------- -----------------------
10 ACCOUNTING NEW YORK 3
20 RESEARCH DALLAS 5
30 SALES CHICAGO 6
NULL is an unknown and not a value, NULL values never satisfy equi-join
condition.

3.1.2. NON EQUI-JOINS (Theta Join)

When the comparison operator used in joining columns is other than
equality, the join is called a Non Equi-join.

Suppose we want to find the grades of employees based on their salary

SQL> select ename,grade,sal from emp,salgrade where sal between

losal and hisal;

Page 50 of 91
DBMS/SQL+

ENAME GRADE SAL

---------------------------------
SMITH 1 800
ADAMS 1 1100
JAMES 1 950
WARD 2 1250
MARTIN 2 1250
MILLER 2 1300
ALLEN 3 1600
TURNER 3 1500
JONES 4 2975
BLAKE 4 2850
CLARK 4 2450
SCOTT 4 3000
FORD 4 3000
KING 5 5000

14 rows selected

3.1.3. SELF JOIN

In a self-join a Cartesian product is taken on two sets of rows of the same
table and a condition is applied on these rows.
.

TABLE F1
COL1 COL2
----- -----
A 1
B 2
C 3

SELECT * FROM f1 first_f1, f1 second_f1

WHERE first_f1.col1 = second_f1.col1;

COL1 COL2 COL1 COL2

--------- -------- --------- ---------
A 1 A 1
B 2 B 2
C 3 C 3

In the above query the join is taken between the rows of the same table so it
also becomes a self-join.

To get the names and salaries of all employees who have managers along with
manager’s name and managers salary.

SQL> SELECT a.ename name , a.sal salary , b.ename mgr, b.sal " Mgr
Salary" FROM emp a , emp b WHERE A.mgr = B.empno;

Page 51 of 91
DBMS/SQL+

NAME SALARY MGR Mgr Salary

---------- ------- -------- ---------------
SMITH 800 FORD 3000
ALLEN 1600 BLAKE 2850
WARD 1250 BLAKE 2850
JONES 2975 KING 5000
MARTIN 1250 BLAKE 2850
BLAKE 2850 KING 5000
CLARK 2450 KING 5000
SCOTT 3000 JONES 2975
TURNER 1500 BLAKE 2850
ADAMS 1100 SCOTT 3000
JAMES 950 BLAKE 2850
FORD 3000 JONES 2975
MILLER 1300 CLARK 2450
13 rows selected.
* King is not selected because he does not have a manager

3.1.4. OUTER JOINS

Consider the query for getting all the department details along with the
number of employees in it. In an equi-join, the details of deptno 40 were
not shown because there are no employees assigned to it in the EMP table.

This can be avoided by taking an outer join. Outer Join is an exclusive union
of sets (whereas normal joins are intersection). Outer Joins can be
simulated using UNIONS.

In a join of two tables an outer join may be for the first table or the second
table. If the outer join is taken on say the dept table then each row of dept
table will be selected at least once whether or not a join condition is
satisfied.
The general syntax is

WHERE table1 < OUTER JOIN INDICATOR > = table 2

In Oracle it is (+)

table1.column = table2.column(+) means OUTER join is taken on

table1.

• To get the details of all departments with number of employees assigned to

it (even 0 employees)
SELECT dept.deptno, dept.dname, dept.loc, count(empno) "No. of
employees" FROM dept, emp

Page 52 of 91
 DBMS/SQL+

WHERE emp.deptno(+)=dept.deptno
GROUP BY dept.deptno,dept.dname,dept.loc;

DEPTNO DNAME LOC No. of employees

----------------------- --------- --------------------
10 ACCOUNTING NEW YORK 3
20 RESEARCH DALLAS 5
30 SALES CHICAGO 6
40 OPERATIONS BOSTON 0

Suppose there were any employees without deptno assigned to it. If we take
the equi-join then employees without any dept will not be selected because join
condition will not be satisfied for NULL values. To solve this problem we have
to take an outer join on the department table.

The query to get all the names, salary, department names of all employees we
have to take an outer join on employee table.

SELECT ename, sal,dname FROM emp,dept WHERE

dept.deptno(+)=emp.deptno;

3.2. Subqueries
When the WHERE clause needs a set of values which can be only obtained from
another query, the subquery is used. In the WHERE clause it can become a part
of the following predicates
• COMPARISON Predicate
• IN Predicate
• ANY or ALL Predicate
• EXISTS Predicate

It can be also used as a part of the condition in the HAVING clause.

To list the names of all employees working in SALES dept, assuming that deptno
of SALES dept is not known.

SELECT ename,sal FROM emp

WHERE deptno = (SELECT deptno
FROM dept
WHERE dname = 'SALES');
ENAME SAL
----------- ----------
ALLEN 1600
WARD 1250
MARTIN 1250

Page 53 of 91
DBMS/SQL+

BLAKE 2850
TURNER 1500
JAMES 950

The first select statement is called as OUTER QUERY or MAIN QUERY. The second
SELECT statement is called INNER QUERY or SUB-QUERY. All the rules, which
apply to the main query, also apply to the sub-query. The inner query is
executed FIRST and ONLY ONCE. The inner query returns a single value, namely
30. This value is used to evaluate the main query. Notice that the table used in
inner query can be different from the table used in outer query.

• The sub-query cannot have an ORDER BY clause and UNION of SELECT’s

• The sub-query may select only one column except in the case of EXISTS
predicate
There are two types of sub-queries – Co-related and Non Co-related sub-queries.
If the sub-query uses any data from the FROM clause of the outer query, the
sub-query is evaluated for each row of the outer query. The sub-query thus has
to be repeated for each row of the outer query and is called co-related sub-
query.

When no data is needed from the outer query, the sub-query is evaluated only
once. The sub-query is executed and the result set is obtained. Such a sub-query
is called as non co-related sub-query.

3.2.1. SUB-QUERIES IN PREDICATES

Comparison
<EXPRESSION> < OPERATOR > < SUBQUERY>

The sub-query should result in one value of the same data type as the left-
hand side.
To get all the employees whose salary is greater than the average salary of
the company.

SQL> SELECT ename,sal FROM emp

WHERE sal > ( SELECT AVG(sal) FROM emp);

ENAME SAL
---------- ----------
CLARK 2450
BLAKE 2850
JONES 2975
SCOTT 3000
FORD 3000
KING 5000

Page 54 of 91
DBMS/SQL+

To select all employees who do the same job as that of 'SCOTT'.

SQL> SELECT empno,ename, sal FROM emp

2 WHERE JOB=(SELECT job FROM emp WHERE ename='SCOTT');

EMPNO ENAME SAL

---------- ---------- ----
7788 SCOTT 3000

3.2.2. IN PREDICATE IN SUB-QUERIES

<EXPRESSION> IN < SUBQUERY>
The list of values generated by the sub-query should be of same datatype as
the left-hand side.
To see all details of departments who have employees working in it.

SQL> SELECT * FROM dept

WHERE deptno IN (SELECT distinct deptno FROM emp);

DEPTNO DNAME LOC

---------- ---------- -------------
10 ACCOUNTING NEW YORK
20 RESEARCH DALLAS
30 SALES CHICAGO

3.2.3. ALL, ANY PREDICATES

These predicates are called Quantified Predicates. They allow testing of a
single value against all members of the set.
The general form is
< EXPRESSION> <OPERATOR> < ANY /ALL/SOME > <SUB-QUERY>

The datatypes must be comparable to the left and right hand side. The sub-
query generates a list of values of the same type as the <EXPRESSION> on the
left and the operator is any comparison operator.

FOR ALL the predicate is evaluated as follows:

1. True if the comparison is true for every value of the list of values.
2. True if sub-query gives a null set (No values)
3. False if the comparison is false for one or more of the list of values
generated by the sub-query.
FOR ANY the predicate is evaluated as follows:
1. True if the comparison is true for one or more values generated by the
sub-query.
2. False if sub-query gives a null set (No values)

Page 55 of 91
DBMS/SQL+

3. False if the comparison is false for every value of the list of values
generated by the sub-query.

Examples of ANY, ALL

To find names and salary details of all employees who earn more than the
managers.

SQL> SELECT ename,sal FROM emp

2 WHERE sal > ANY ( SELECT sal FROM emp WHERE JOB=’MANAGER’) AND JOB
!=’MANAGER’;
ENAME SAL JOB
---------- --------- ---------
SCOTT 3000 ANALYST
KING 5000 PRESIDENT
FORD 3000 ANALYST

3 rows selected.
To select the employees whose sal is greater than the sal of all the employees
working in deptno 30.

SQL> SELECT ename,sal FROM emp

2 WHERE sal > ALL ( SELECT sal FROM emp WHERE deptno=30);

ENAME SAL
------------ ----------
JONES 2975
SCOTT 3000
KING 5000
FORD 3000

3.2.4. Co- Related Sub-queries

In the inner query one may refer to the identifiers available from the outer
query with proper identification.

SQL> SELECT ename, mgr, hiredate FROM emp a

2 WHERE a.hiredate < (SELECT b.hiredate FROM emp b
WHERE b.empno=a.mgr);

ENAME MGR HIREDATE

---------- ------- ----------------
SMITH 7902 17-DEC-80
ALLEN 7698 20-FEB-81
WARD 7698 22-FEB-81
JONES 7839 02-APR-81
BLAKE 7839 01-MAY-81
CLARK 7839 09-JUN-81
6 rows selected.

Page 56 of 91
DBMS/SQL+

Here two sets of rows of EMP table will be created one for the outer query and
one for the inner query. In the first pass all the rows of the second set will be
compared with first row of first set. This process will be repeated for all the
rows of the first set. If for each comparison the condition becomes true then
the row from the first set will be selected.

SELECT ename, sal, job from emp a

Where sal > (select avg (sal) from emp b where b.job = a.job);

ENAME SAL JOB

SMITH 1600 SALESMAN

JONES 2975 MANAGER
BLAKE 2850 MANAGER
TURNER 1500 SALESMAN
MILLER 1300 CLERK

3.2.5. EXISTS Predicate

The EXISTS Predicate is of the form

<Query> WHERE
EXISTS < sub-query>
Query will be evaluated if EXISTS takes a value TRUE. It takes the Value TRUE if
the <sub-query> result table is non null and FALSE if it is Null. It is mostly used
with Co-Related Subqueries. EXISTS does not check for a particular value. It
checks whether subquery returns rows or not which match to outer query
rows.

To list all details of dept which has at east one employee assigned to it.

SQL> SELECT * FROM dept a

2 WHERE EXISTS (SELECT * FROM EMP b
WHERE b.deptno = a.deptno);
DEPTNO DNAME LOC
------ -------------- -------------
10 ACCOUNTING NEW YORK
20 RESEARCH DALLAS
30 SALES CHICAGO

Here there are 4 rows in dept table so the Inner Query will be executed 4 times
for each row returned by outer query. Since there are no employees for deptno
40 the sub-query will return no rows i.e. a null set. Hence EXISTS will become
false and the outer query will not select the row in dept table with value 40.

Page 57 of 91
 DBMS/SQL+

4. Advanced SQL
4.1. Index
Index is a database object, which speeds up access to the data. There are
different types of index, which the user can create.
CREATE [UNIQUE] INDEX index_name
ON table_name(col_name1 [ASC|DESC],col_name2,.....);

• A table can have any number of indices

• An index can be on a single column or on multiple columns. In 8i, Up to
32 columns can be included in one index
• Updation of all indices is handled by Oracle
• UNIQUE ensures that all values in an index are unique
• Index is ascending by default
• ORACLE decides when to use the index while accessing the data
• Removal of an index does not affect the table on which it is based
• When you create a primary or a unique constraint on the table, a unique
index is automatically created for you. The name of the constraint will
be used for the index name

Example :
CREATE INDEX emp_sal_index ON emp(sal);

• To allow only unique values in ename field, the create statement will
be,

CREATE UNIQUE INDEX emp_ename_unindex ON emp( ename );

4.2. Synonym
Synonyms are objects, which contain a duplicate name of an existing object.

CREATE [PUBLIC] SYNONYM another_name

FOR existing_name

• Existing_name can be name of a table or view or sequence

• PUBLIC is used to grant permission to all users to access the object using
the new name. (Only by DBA)
• Useful in hiding ownership details of an object

CREATE PUBLIC SYNONYM new_emp FOR emp;

Page 58 of 91
 DBMS/SQL+

4.3. Sequence
Sequence is an object, which can be used to generate sequential numbers.
The numbers, which are generated using a sequence, are usually used to fill
up columns that are declared as UNIQUE or PRIMARY KEY columns.

CREATE SEQUENCE seq_name

[INCREMENT BY n1]
[START WITH n2]
[MAXVALUE n3]
[MINVALUE n4]
[CYCLE|NOCYCLE];

• START WITH indicates the first number in the series

• INCREMENT BY is the difference between consecutive numbers. If n1 is
negative, numbers that are generated are descending in order
• MAXVALUE & MINVALUE indicate the extreme values
• CYCLE indicates that once the extreme is reached, start the cycle again
with n2
• NOCYCLE means that once the extreme is reached stop generating
numbers

CREATE SEQUENCE s1
INCREMENT BY 1
START WITH 1
MAXVALUE 10000
NOCYCLE ;
s1 will generate NUMBERs 1,2,3....10000 and stop.

The NUMBER generated by s1 can be used to fill empno column of emp table.

4.4. View
View is a logical table based on one or more tables.

CREATE VIEW view_name [alias]

AS query
[WITH CHECK OPTION];

• View can be used as if it is a table

• View does not contain data
• Whenever a view is accessed, the query is evaluated. Thus view is
dynamic
• Any changes made in the view affect the tables on which the view is
based

Page 59 of 91
DBMS/SQL+

• View helps to hide the ownership details of a table and complexity of

query used to retrieve data, from the user
• "With check option" means you cannot insert a row in the underlying
table, which cannot be selected using the view
• If you use a order by clause in the view, then it becomes a read only
view automatically

CREATE VIEW newempview

AS
SELECT * FROM emp
WHERE hiredate >'01-jan-82';

Read Only View

• If a view is based on a single table the user may manipulate records in
the view and its underlying base table. The WITH READ ONLY clause of
the CREATE VIEW command can be used to prevent users from
manipulating records in a view.

There are certain restrictions on queries, which can be used to create a view. The
restrictions will be explained when queries are dealt with.

Updation /Insertion of rows in the base table using views is possible only if the
view is based on a single table with some restrictions. The restrictions are:

1. Updation /Insertion Not possible if view is based on two tables. Can be done
in ORACLE 8 by using INSTEAD OF trigger.
2. Insertion is not allowed if the underlying table has any not NULL columns
that do not appear in the view.
3. Insertion/Updation is not allowed if any column of the view referenced in
update/Insert contains functions or calculations.
4. Insertion/Updation /deletion not allowed if view contains group by or
distinct clause in the query.

Page 60 of 91
 DBMS/SQL+

5. PL/SQL
5.1. Introduction to PL/SQL
Procedural Language (PL/SQL) is a procedural extension to SQL, provided by
ORACLE. Through PL/SQL, the data manipulation capabilities of SQL are
combined with the processing capabilities of a procedural language. (Pro * C,
Pro * COBOL).

PL/SQL provides features like conditional execution, looping and branching.

PL/SQL supports subroutines too.

The basic unit in any PL/SQL program is a block, which can occur
sequentially (one after the other) or nested (One inside the other).

Anonymous blocks are generally constructed dynamically and executed only

once.

Named blocks are anonymous blocks with a label that gives the block a
name. They are also constructed dynamically and executed only once.

Subprograms are procedures, packages and functions that are stored in the
database. These blocks generally don’t change once they are constructed
and are executed many times by calling them explicitly via a call to a
procedure, package or function.

TRIGGERS are named blocks that are stored in the database. They are
executed implicitly whenever the triggering event occurs. The triggering
event is a DML statement executed against a table in the database.

5.2. PL/SQL block and its Sections

A PL/SQL block is a set of SQL statements, which collectively accomplish a
particular task. SQLPLUS commands cannot be used in a PL/SQL block. DDL
commands such as CREATE and ALTER cannot be used in a PL/SQL block. All
DML and DCL statements can be used in PL/SQL though some modification is
required in the syntax of some of the commands.

A PL/SQL block is called so, as the statements are written in a block-

structured way as in PASCAL. Every block has a keyword, which indicates the

Page 61 of 91
 DBMS/SQL+

beginning of a block, and optionally a keyword, which indicates the end of a

block.

SAMPLE PL/SQL BLOCK

DECLARE -- Declaration Section
V_Salary NUMBER(7,2);
/* V_Salary is a variable declared in a PL/SQL block. This variable
is used to store JONES' salary. */
BEGIN -- Execution Section
SELECT sal INTO V_Salary
FROM emp WHERE ename = 'JONES'
EXCEPTION -- Exception Section

WHEN no_data_found THEN

DBMS_OUTPUT.PUT_LINE(‘Employee Jones does not exist’);
END ; -- End of Block
/ -- PL/SQL block terminator

Output
SQL> /
PL/SQL procedure successfully completed.

Example 5.1

Notations are as follows:

1. -- is a single line comment.
2. /* */ is a multi-line comment.
3. Every statement must be terminated by a semicolon.
4. PL/SQL block is terminated by a slash / on a line by itself

A PL/SQL block must have a Declaration section and an Execution section. It

can optionally have an Exception section too.

Declaration section is where all the variables and cursors used in a block are
declared. Execution section has SQL statements and PL/SQL constructs which
aid in BRANCHING, LOOPING and CONDITIONAL EXECUTION. Exception
section has statements, which handle the exceptions.

5.3. Declaration Section

Declaration section is used to declare variables, constants, exceptions and
cursors. PL/SQL supports all ANSI SQL datatypes.

Page 62 of 91
DBMS/SQL+

5.3.1. Variable Declaration

Communication with the database takes place via variables in the PL/SQL
block.
The general syntax for declaring a variable is

Variable_Name TYPE [CONSTANT] [DEFAULT] [ NOT NULL] [:= Value];

DECLARE
V_Description VARCHAR2(25) ;
V_Sal NUMBER(5) NOT NULL := 3000 ;
V_Compcode VARCHAR2(20) CONSTANT := 'ABC CONSULTANTS' ;
V_Comm NOT NULL DEFAULT 0 ;

If a variable is not initialized, it is assigned a NULL value by default. If a

NOT NULL is present in the declaration then the variable must be initialized.
You cannot assign a null value to a NOT NULL variable in the executable or
exception section.

If a CONSTANT is present in the declaration, the variable must be initialized

and its value cannot be changed from its initial value.

5.3.2. PL/SQL Datatypes

All PL/SQL datatypes are classified as scalar, reference and composite type.
Scalar datatypes do not have any components within it, while composite
datatypes have other datatypes within them. A reference datatype is a
pointer to another datatype.

PL/SQL Datatypes

Scalar type Reference type Composite type

(number,varchar2,Boolean (tablename.columnname%type (pl/sql record
type
binary_integer etc.) tablename%rowtype) pl/sql table
type)

Figure 5-1 Datatypes

5.3.3. SCALAR TYPES

1. NUMBER

Page 63 of 91
DBMS/SQL+

This can hold a numeric value, either integer or floating point. It is same
as the number database type.
2. BINARY_INTEGER
If a numeric value is not to be stored in the database, the
BINARY_INTEGER datatype can be used. It can only store integers from -
2147483647 to + 2147483647. Mostly used for counter variables.
V_Counter BINARY_INTEGER DEFAULT 0 ;
3. VARCHAR2(L)
L is necessary and is max length of the variable. This behaves like
VARCHAR2 database type. The maximum length in PL/SQL is 32,767
bytes where as VARCHAR2 database type can hold max 2000 bytes. If a
VARCHAR2 PL/SQL column is more than 2000 bytes, it can only be
inserted into a database column of type LONG .
4. CHAR(L)
Here L is the maximum length. Specifying length is optional. If not
specified it defaults to 1. The maximum length of CHAR PL/SQL variable
is 32,767 bytes whereas that of database CHAR column is 255 bytes.
Therefore a CHAR variable of more than 255 bytes can be inserted in the
database column of VARCHAR2 or LONG type.
5. LONG
PL/SQL LONG type is just 32,767 bytes. It behaves similar to LONG
DATABASE type.
6. DATE
The DATE PL/SQL type behaves the same way as the date database
type. The DATE type is used to store both date and time. A DATE
variable is 7 bytes in PL/SQL.
7. BOOLEAN
A Boolean type variable can only have one of 2 values either TRUE or
FALSE. They are mostly used in control structures.
V_Does_Dept_Exist BOOLEAN := TRUE ;
V_ Flag BOOLEAN := 0 ; -- illegal

5.3.4. REFERENCE TYPES

A reference type in PL/SQL is the same as a pointer in C. A reference type
variable can point to different storage locations over the life of the program.

Page 64 of 91
 DBMS/SQL+

USING %TYPE
%TYPE is used to declare a variable with the same datatype as a column of a
specific table.
Var_Name table_name.col_name%TYPE ;
V_Empno emp.empno%TYPE ;
Datatype of V_Empno is same as that of Empno column of the EMP table.

USING %ROWTYPE
%ROWTYPE is used to declare a compound variable whose type is same as that
of a row of a table.
Var_Name table_name%ROWTYPE ;
V_Emprec emp%ROWTYPE ;

V_Emprec is a variable, which contains within itself as many variables whose

names and datatypes match those of the EMP table. To access the Empno
element of V_Emprec, use
V_Emprec.empno ;

5.4. Variable Scope And Visibility

DECLARE
V_Flag BOOLEAN ;
V_Var1 CHAR(9);
BEGIN
1 V_Flag and the CHAR(9) V_Var1 are visible
DECLARE
V_Var1 NUMBER(9);
V_Date DATE;
BEGIN
2 V_Flag, V_Date and the NUMBER(9) V_Var1 are visible
END;

3 V_Flag and the CHAR (9) V_Var1 are visible

END;

Page 65 of 91
DBMS/SQL+

<<Outer>>
DECLARE
v_NAME VARCHAR2(10);
v_SSN NUMBER(9);
BEGIN V_NAME and the NUMBER(9) v_SSN are visible
À
DECLARE
v_SSN CHAR(11);
v_FLAG VARCHAR2(10); v_Flag, and the CHAR(11) v_SSN are visible. But
BEGIN we can refer to the NUMBER(9)v_SSN with
Á IOuter.v_SSN
END;
 V_NAME and the NUMBER(9) v_SSN are visible
END;

SQL IN PL/SQL
SELECT STATEMENT
SELECT Column_List INTO Variable_List
FROM Table_List
[WHERE expr1]
[CONNECT BY expr2 [START WITH expr3]]
[GROUP BY expr4] [HAVING expr5]
[UNION | INTERSECT | MINUS SELECT ...]
[ORDER BY expr | ASC | DESC]
[FOR UPDATE [OF Col1,...] [NOWAIT]]
INTO Variable_List

The column values returned by the SELECT command must be stored in

variables. The Variable_List should match Column_List in both count and
datatype. Here the variable_list is PL/SQL (Host) variables, which should be
defined before use.

This clause is used if the selected row must be modified through a DELETE or
UPDATE command. Since the contents of the row must not be modified between
the SELECT command and the UPDATE command, it is necessary to lock the row
after the SELECT command. FOR UPDATE will lock the selected row. OF Col1,
lists the columns, which can be modified by the UPDATE command. If OF
Col1,... is missing, all the columns can be modified.

It is possible that when SELECT..UPDATE is issued, the concerned row is already

locked by some other user or a previous SELECT..UPDATE command. If NOWAIT
is not given, then the current SELECT..UPDATE command will wait until the lock

Page 66 of 91
DBMS/SQL+

is cleared. IF NOWAIT is given, then the SELECT...UPDATE will return

immediately with a failure.

DECLARE
V_Emprec emp%ROWTYPE;
V_Ename emp.ename%TYPE;
V_Sal emp.sal%TYPE;
BEGIN
SELECT * INTO V_Emprec FROM emp WHERE ename ='SMITH';
SELECT ename,sal INTO V_Ename,V_Sal from emp where ename ='SMITH';
END;

Example 5.2
INSERT STATEMENT
The syntax for the insert statement remains the same as in SQL –INSERT.

DECLARE
V_Dname VARCHAR2(15) := 'FINANCE';
BEGIN
INSERT INTO dept VALUES ( 50 , V_Dname, 'BOMBAY');
END;
/

Example 5.3
DELETE STATEMENT
DELETE FROM table_name
WHERE [CURRENT OF Cursor_Name|Condition]

DECLARE
V_Sal_Cutoff NUMBER(4):= 2000;
BEGIN
DELETE FROM emp WHERE sal < V_Sal_Cutoff;
END;
/

Example 5.4
UPDATE STATEMENT
UPDATE table_name SET col_name = Value
WHERE [CURRENT OF Cursor_Name|Condition]
DECLARE
V_Sal_Incr NUMBER(5) := 1000;
BEGIN
UPDATE emp SET sal = sal + V_Sal_Incr WHERE ename='SMITH';
END;
/
Example 5.5

Page 67 of 91
DBMS/SQL+

VARIABLE NAMES
Consider the block

DECLARE
V_Deptno NUMBER(10) := 30;
V_Dname VARCHAR2(15) ;
BEGIN
SELECT dname INTO V_Dname FROM dept WHERE deptno = V_Deptno;
END;

Example 5.6
Here there is no ambiguity between the database column names and PL/SQL
variables.

DECLARE
Deptno NUMBER(10) := 30;
V_dname VARCHAR2(15) ;
BEGIN
SELECT dname INTO V_dname FROM dept WHERE deptno = deptno;
DELETE FROM dept WHERE deptno = deptno ;
END;

Example 5.7
Here the select statement will select names of all the departments and not only
deptno. 30. The delete statement will delete all the employees. This happens
because when the PL/SQL engine sees a condition expr1 = expr2, expr1 and
expr2 are first checked to see whether they match the database columns first
and then the PL/SQL variables. So in above example deptno = deptno both are
treated as database columns and the condition will become true for every row
of the table.

If a block has a label then variables with same names as database columns can
be used by using <<blockname>>. Variable_Name notation.

<<BLOCK1>>
DECLARE
Deptno NUMBER(10) := 30;
Dname VARCHAR2(15) ;
BEGIN
SELECT Dname INTO dname FROM dept WHERE deptno = Block1. deptno;
DELETE FROM dept WHERE deptno = Block1. deptno ;
END;

Example 5.8

Page 68 of 91
 DBMS/SQL+

Although this method can be used to get the desired results, it is not a good
programming practice to use same names for PL/SQL variables and database
columns.

5.5. Programming Constructs

5.5.1. LOOPING
A LOOP is used to execute a set of statements more than once. The syntax of
a LOOP statement is
LOOP
PL/SQL_Statements;
END LOOP ;

DECLARE
V_Counter NUMBER := 50 ;
BEGIN
LOOP
INSERT INTO dept
VALUES(V_Counter,'NEW DEPT','SOMEWHERE');
V_Counter := V_Counter + 10 ;
END LOOP;
COMMIT ;
END ;
/

Example 5.13

This is an endless loop. When LOOP ENDLOOP is used in the above format,
then an exit path must necessarily be provided. Exit path is provided using
EXIT or EXIT WHEN commands.

EXIT is an unconditional exit. Control is transferred to the statement

following END LOOP, when the execution flow reaches the EXIT statement.

BEGIN
.....
LOOP
IF <Condition> THEN
. . . . .
EXIT ; -- Exits loop immediately
END IF ;
END LOOP;
. . . . . -- Control resumes here
COMMIT ;
END ;
/

Page 69 of 91
DBMS/SQL+

EXIT WHEN is for conditional exit out of the loop.

DECLARE
V_Counter NUMBER := 50 ;
BEGIN
LOOP
INSERT INTO dept
VALUES(V_Counter,'NEWDEPT','SOMEWHERE') ;
DELETE FROM emp WHERE deptno = V_Counter ;
V_Counter := V_Counter + 10 ;
EXIT WHEN V_Counter >100 ;
END LOOP;
COMMIT ;
END ;
/

Example 5.14
As long as V_Counter has a value less than or equal to 100, the loop
continues.

LOOP…END LOOP can be used in conjunction with FOR and WHILE for better
control on looping.
The syntax of FOR LOOP is

FOR Variable IN [REVERSE] Lower_Bound..Upper_Bound

LOOP
PL/SQL_Statements
END LOOP ;

FOR loop is used for executing the loop a fixed number of times. The number
of times the loop will execute equals
Upper_Bound - Lower_Bound + 1.
Upper_Bound and Lower_Bound must be integers and Upper_Bound must be
equal to or greater than Lower_Bound. Variables in FOR loop need not be
explicitly declared. Variables take values starting at a Lower_Bound and
ending at a Upper_Bound. The variable value is incremented by 1, every
time the loop reaches the bottom. When the variable value becomes equal
to the Upper_Bound then the loop executes and exits.

When REVERSE is used, then the variable takes values starting at

Upper_Bound and ending at Lower_Bound. Value of the variable is
decremented each time the loop reaches the bottom.

Page 70 of 91
DBMS/SQL+

DECLARE
V_Counter NUMBER := 50 ;
BEGIN
FOR Loop_Counter IN 2..5
LOOP
INSERT INTO dept
VALUES(V_Counter,'NEW DEPT','SOMEWHERE') ;
V_Counter := V_Counter + 10 ;
END LOOP;
COMMIT ;
END ;

Example 5.15
The loop will be executed (5 - 2 + 1) = 4 times. A Loop_Counter variable can
also be used inside the loop, if need be. Lower_Bound and/or Upper_Bound
can be integer expressions also.
The syntax for the WHILE LOOP is

WHILE Condition
LOOP
PL/SQL Statements;
END LOOP;
EXIT OR EXIT WHEN can be used inside the while loop to exit the loop
prematurely.

5.5.2. LABELING LOOPS

Loops themselves can be labeled as in the case of blocks. The label can be
used with the EXIT statement to exit out of a particular loop

BEGIN
<<Outer_Loop>>
LOOP
PL/SQL
<< Inner_Loop>>
LOOP
PL/SQL Statements ;
EXIT Outer_Loop WHEN <Condition Met>;
END LOOP Inner_Loop;
END LOOP Outer_Loop;
END ;
/

Example 5.16

Page 71 of 91
DBMS/SQL+

5.5.3. CONDITIONAL EXECUTION

Conditional Execution construct is used to execute a set of statements only
if a particular condition is TRUE or FALSE.

IF Condition_Expr THEN
PL/SQL_Statements
End if;

When the condition evaluates to TRUE then the PL/SQL statements are
executed, otherwise the statement following END IF is executed.

IF V_Empno = 7902 THEN

UPDATE emp
SET sal = sal + 100 WHERE empno = 7902;
END IF;

UPDATE statement is executed only if value of V_Empno variable equals

7902.
PL/SQL allows many variations of the IF – END IF construct.
To take alternate action if the condition is FALSE, use

IF Condition_Expr THEN
PL/SQL_Statements_1 ;
ELSE
PL/SQL_Statements_2 ;
END IF;

When the condition evaluates to TRUE PL/SQL_Statements_1 is executed,

otherwise PL/SQL_Statements_2 is executed.

The above syntax checks only one Condition_Expr.

To check for multiple conditions and take action based on the condition, use

IF Condition_Expr_1 THEN
PL/SQL_Statements_1 ;
ELSIF Condition_Expr_2 THEN
PL/SQL_Statements_2 ;
ELSIF Condition_Expr_3 THEN
PL/SQL_Statements_3 ;
ELSE
PL/SQL_Statements_n ;
END IF;

Page 72 of 91
DBMS/SQL+

As every condition must have at least one statement, NULL statement can be
used as filler. NULL command does nothing. Sometimes NULL is used in a
condition merely to indicate that such a condition has also been taken into
consideration. So your code will resemble the code as given below:

IF Condition_Expr_1 THEN
PL/SQL_Statements_1 ;
ELSIF Condition_Expr_2 THEN
PL/SQL_Statements_2 ;
ELSIF Condition_Expr_3 THEN
Null;
END IF;

CONDITIONS FOR NULL are checked through IS NULL and IS NOT NULL
predicates.

Page 73 of 91
 DBMS/SQL+

6. Error Handling
A good programming language should provide capabilities of handling errors and
recovering from them if possible. PL/SQL implements error handling via
exceptions and exception handlers.

6.1. Types of Errors in PL/SQL

Compile Time errors are reported by the PL/SQL compiler and you have to
correct them before recompiling.
The run-time engine reports run time errors and they are handled
programmatically by raising an exception and catching it in the Exception
section

6.2. Declaring Exceptions

Exceptions are declared in the Declaration section, raised in the executable
section and handled in the Exception Section. There are two types of
exceptions predefined and user defined.

6.2.1. User Defined Exceptions

It is an error defined by the program. It could be an error with the data.

DECLARE
E_Balance_Not_Sufficient EXCEPTION;
E_Comm_Too_Large EXCEPTION;
…
BEGIN
NULL;
END;

6.2.2. Predefined Exceptions

ORACLE has predefined several exceptions that correspond to the most
common ORACLE errors. They are always available to the program; hence
there is no need to declare them. These exceptions are automatically raised
by ORACLE whenever that particular error condition occurs.

Some of the Predefined Exceptions

NO_DATA_FOUND
This is raised when SELECT INTO …. Statement does not return any rows.

Page 74 of 91
DBMS/SQL+

TOO_MANY_ROWS
This is raised when SELECT INTO …. Statement returns more than one row.

INVALID_CURSOR
This is raised when an illegal cursor operation is performed such as closing an
already closed cursor.

VALUE_ERROR
This exception is raised when an arithmetic, conversion, truncation or
constraint error occurs in a procedural statement.

DUP_VAL_ON_INDEX
This is raised when the UNIQUE CONSTRAINT is violated.

Raising Exceptions
When the error associated with an exception occurs, the exception is raised.
This is done through the RAISE command. The syntax is

RAISE Exception_Name ;
An exception is defined and raised as shown below:

DECLARE
...
Retired_Emp EXCEPTION ;
BEGIN
PL/SQL_Statements ;
IF Error_Condition THEN
RAISE Retired_Emp ;
PL/SQL_Statements ;
END IF;
EXCEPTION
WHEN Retired_Emp THEN
PL/SQL_Statements ;
END ;

Control passing to Exception Handler

Page 75 of 91
DBMS/SQL+

DECLARE
Dup_Deptno EXCEPTION;
V_Counter BINARY_INTEGER;
V_Department NUMBER(2) := 50;
BEGIN
SELECT COUNT(*) INTO V_Counter FROM Dept WHERE deptno = 50;
IF V_Counter > 0 THEN
RAISE Dup_Deptno ;
END IF;
INSERT INTO dept VALUES (V_Department ,'NEW NAME', 'NEW LOC');
EXCEPTION
WHEN Dup_Deptno THEN
INSERT INTO Error_Log VALUES ('Department ' || V_Department
|| ' IS ALREADY PRESENT');
END ;
/

Example 6.1
OTHERS exception handler
The OTHERS handler will execute for all raised exceptions. It should always be
the last handler in the block. The OTHERS exception handler handles all errors,
which are not handled separately in the EXCEPTION section. To handle a
specific case within the OTHERS handler, predefined functions SQLCODE and
SQLERRM are used.
SQLCODE returns the current error code. And SQLERRM returns the current
error message text. The values of SQLCODE and SQLERRM should be assigned to
local variables before using it within a SQL statement.

DECLARE
Err_Num NUMBER ;
Err_Msg CHAR(100);
BEGIN
-------
------
EXCEPTION
----------
-------
WHEN OTHERS THEN
-- Assign values to variables. SQLCODE SQLERRM cannot be used
directly.
Err_Num = SQLCODE ;
Err_Msg =SUBSTR( SQLERRM, 1, 100);
-- Errors is a table existing in the database created by the user
INSERT INTO errors VALUES( Err_Num, Err_Msg );
END;
/

Page 76 of 91
DBMS/SQL+

DECLARE
V_Deptno NUMBER(2):= 10;
BEGIN
DELETE FROM dept WHERE deptno = V_Deptno ;
EXCEPTION
WHEN OTHERS THEN
/* Error -2292 OCCURS WHEN REFERENTIAL INTEGRITY RULE
IS VIOLATED */
IF SQLCODE = - 2292 THEN
/* Error_Log is a table containing one varchar2 column
created by user to store error messages */
INSERT INTO Error_Log
VALUES('Employees exist for Department No' || V_Deptno);
ELSE
INSERT INTO Error_Log
VALUES('Unable to delete dept No' || V_Deptno ||'because
of unknown reasons');
END IF ;
END ;
/
DECLARE
V_Dummy VARCHAR2(1);
V_Department NUMBER(2) := 50;
BEGIN
SELECT 'X' INTO V_Dummy FROM dept WHERE Deptno= V_Department;
INSERT INTO Error_Log
VALUES ('Department ' || V_Department || 'IS ALREADY PRESENT');
EXCEPTION
WHEN NO_DATA_FOUND THEN
INSERT INTO dept
VALUES (V_Department ,'NEW NAME', 'NEW LOC');
END ;
/
Example 6.2

Page 77 of 91
 DBMS/SQL+

7. Stored Procedures and Functions

The PL/SQL blocks in the examples discussed so far do not have any name or
identity whatsoever. Even though, we have named some of the loops, we have
not been able to give a name to the PL/SQL block as a whole. Hence such blocks
are known as Anonymous Blocks.

Anonymous blocks are executed by interactively entering the block at the SQL>
prompt OR by writing the PL/SQL statements in a user-named file and executing
the block at SQL> prompt using @ command. The block needs to be compiled
every time it is run and only the user who created the block can use the block.

Stored subprograms are named PL/SQL blocks. Stored subprograms are compiled
at the time of creation and stored in the database itself. The source code is also
stored in the database. Any user with necessary privileges can use the stored
subprogram. Stored subprogram come in handy during application development.

Stored subprograms can be executed from the SQL> prompt using EXECUTE
command. The syntax is

SQL> EXECUTE Prg_Name(Parameter_List)

Stored subprograms can also be invoked from another PL/SQL block, similar to
calling a function in C.

Use SHOW ERRORS to see the Compile Time errors.

Use USER_OBJECTS, USER_ERRORS and USER_SOURCE to know about names,

errors and Source code of stored subprograms.

7.1. Procedure
CREATE PROCEDURE Proc_Name
(Parameter {IN | OUT | IN OUT} datatype := value,...) AS
Variable_Declaration ;
Cursor_Declaration ;
Exception_Declaration ;

BEGIN
PL/SQL_Statements ;
EXCEPTION
Exception_Definition ;
END Proc_Name ;

Page 78 of 91
DBMS/SQL+

In a procedure declaration, it is illegal to constraint CHAR and VARCHAR2

parameters with length and NUMBER parameters with precision and scale.

If a parameter is defined as IN, then the calling program can only pass the value
to the procedure. It cannot be on the left-hand side of the assignment operator.

A parameter defined as OUT is used to return a value to the caller. Any number
of OUT parameters can be passed. Any OUT parameter cannot be on the right
hand side of the assignment operator.

A parameter declared as IN OUT is used to pass an initial value to the procedure

and on termination of the procedure, caller gets the current of value of the
parameter. Any change in the value of an IN OUT parameter inside the
procedure is reflected as the current value of the parameter.

IN OUT IN OUT

the default must be specified must be specified

used to pass values to used to return values to Used to pass intial
the procedure. the caller. values to the procedure
and return updated
values to the caller.
formal parameter acts formal parameter acts formal parameter acts
like a constant. like an uninitialized like an initialized
variable. variable..
formal parameter formal parameter formal parameter should
cannot be assigned a cannot be used in an be assigned a value.
value. expression, but should
be assigned a value.
actual parameter can actual parameter must actual parameter must
be a constant, literal, be a variable. be a variable.
initialized variable or
expression.

Page 79 of 91
DBMS/SQL+

CREATE OR REPLACE PROCEDURE

Raise_Salary ( P_Empno IN NUMBER, p_raise in NUMBER) IS
V_Cur_Salary NUMBER ;
Missing_Salary EXCEPTION;
BEGIN
-- V_Cur_Salary is used only to fetch the data
SELECT sal into V_Cur_Salary FROM emp WHERE empno=P_Empno;
IF V_Cur_Salary IS NULL THEN
RAISE Missing_Salary;
END IF ;
UPDATE emp SET sal = sal + p_raise
WHERE empno = P_Empno ;
EXCEPTION
WHEN NO_DATA_FOUND THEN
INSERT INTO Emp_Audit
VALUES( P_Empno, 'No Employee with id ' || P_Empno);
WHEN Missing_Salary THEN
INSERT INTO Emp_Audit
VALUES( P_Empno, 'SALARY IS MISSING');
END;
/
CREATE OR REPLACE PROCEDURE
Get_Emp_Details( P_Empno IN NUMBER, P_Ename OUT VARCHAR2,
P_Sal OUT NUMBER ) IS
BEGIN
SELECT ename, sal into P_Ename, P_Sal from emp
WHERE empno=P_Empno;
EXCEPTION
WHEN NO_DATA_FOUND THEN
INSERT INTO Emp_Audit
VALUES( P_Empno, 'No Employee with id ' || P_Empno);
P_Ename := NULL;
P_Sal := NULL;
END ;
/
Example 7.1
Create two bind variables Salary and Name in SQLPLUS using VARIABLE
command as follows
SQL> VARIABLE Salary NUMBER
SQL> VARIABLE Name VARCHAR2(20)
Execute the procedure using the command
EXECUTE Get_Emp_Details(7566,:Salary, :Name)
After execution use the SQL*PLUS PRINT command to see the results

SQL> PRINT Salary

PRINT Name

Page 80 of 91
DBMS/SQL+

7.1.1. PARAMETER DEFAULT VALUES

Like variable declarations, the formal parameters to a procedure or function
can have default values. If a parameter has default values, it does not have
to be passed from the calling environment. If it is passed actual parameter
will be used instead of default. Only IN parameters can have default values.

PROCEDURE Create_Dept( New_Deptno IN NUMBER ,

New_Dname IN VARCHAR2 DEFAULT 'TEMP'
New_Loc IN VARCHAR2 DEFAULT 'TEMP' ) IS
BEGIN
INSERT INTO dept
VALUES ( New_Deptno, New_Dname, New_Loc) ;
END ;

Example 7.2
Now consider the following calls to Create_Dept

BEGIN
Create_Dept( 50);
-- Actual call will be Create_Dept ( 50, 'TEMP', 'TEMP')
Create_Dept ( 50, 'FINANCE');
-- Actual call will be Create_Dept ( 50, 'FINANCE' ,'TEMP')
Create_Dept( 50, 'FINANCE', 'BOMBAY') ;
-- Actual call will be Create_Dept(50, 'FINANCE', 'BOMBAY' )
END;

CREATE OR REPLACE PROCEDURE Insdept

(depno dept.deptno%TYPE,dname dept.dname%TYOE,loc dept.loc%TYPE)
AS
dno dept.deptno%TYPE;
BEGIN
SELECT deptno INTO dno FROM dept WHERE dept.deptno=depno;
DBMS_OUTPUT.PUT_LINE('DUPLICATE VALUE'||DNO);
EXCEPTION
WHEN NO_DATA_FOUND THEN
DBMS_OUTPUT.PUT_LINE('IN NO DATA FOUND');
INSERT INTO dept VALUES(depno,dname,loc);

END;

Example 7.4

Page 81 of 91
 DBMS/SQL+

CREATE OR REPLACE PROCEDURE findComm

(amt in out emp.sal%type)
AS
BEGIN
SELECT sal INTO amt FROM emp WHERE comm=amt;
END;

-- Execution of the procedure findComm

DECLARE
var1 emp.sal%type;
BEGIN
var1 :=300;
findComm(var1);
DBMS_OUTPUT.PUT_LINE('Commission in exec'|| var1);
END;

Example 7.5

7.2. Functions
Functions are similar to procedures. They can accept one or more
parameters and return a single value by using a return value. Functions can
return multiple values by using OUT parameters. Functions are used as part
of an expression and can be called as
Lvalue =Function_Name( Param1, Param2, …….)

Functions returning a single value for a row can be used with SQL
statements.

Syntax of a FUNCTION:

CREATE FUNCTION Func_Name(Param datatype := value,..)

RETURN datatype1 AS
Variable_Declaration ;
Cursor_Declaration ;
Exception_Declaration ;
BEGIN
PL/SQL_Statements ;
RETURN Variable_Or_Value_Of_Type_Datatype1 ;
EXCEPTION
Exception_Definition ;
END Func_Name ;

Write a function, which can be used to insert a new department in dept

table. Function should return the status of the insert operation.

Page 82 of 91
DBMS/SQL+

CREATE FUNCTION Crt_Dept(Dno NUMBER, Dname VARCHAR2, Dloc

VARCHAR2)
RETURN BOOLEAN AS
BEGIN
INSERT INTO dept
VALUES(Dno,Dname,Dloc) ;
RETURN TRUE ;
EXCEPTION
WHEN OTHERS THEN
RETURN FALSE ;
END Crt_Dept ;
/
Example 7.6
Function to calculate average salary of a department.
Returns Average Salary of a department.
Returns -1 if no employees are there in the department.
Returns -2 if any other error.

CREATE OR REPLACE FUNCTION Get_Avg_Sal(P_Deptno IN NUMBER)

RETURN NUMBER AS
V_Sal NUMBER;
BEGIN
SELECT Trunc(Avg(sal)) INTO V_Sal from emp
where deptno=P_Deptno;
IF V_Sal IS NULL THEN
V_Sal := -1 ;
END IF;
RETURN V_Sal ;
EXCEPTION
WHEN OTHERS THEN
RETURN -2; --Signifies any other errors
END Get_Avg_Sal ;

Example 7.7
Create a bind variable Avgsalary in SQLPLUS using VARIABLE command as
follows
SQL> VARIABLE Avgsalary NUMBER
Execute the Function by executing the following PL/SQL block.

Page 83 of 91
 DBMS/SQL+

BEGIN
:Avgsalary := Get_Avg_Sal ( 50);
END ;
/

After execution use the SQL*PLUS PRINT command to see the results
PRINT Avgsalary

Exceptions raised inside Procedures and Functions

If an error occurs inside a procedure, an exception (predefined or user
defined) is raised. If the procedure has no exception handler for this error,
control immediately passes out of the procedure to the calling environment
in accordance with the propagation rules. The values of OUT and IN OUT
formal parameters are not returned to actual parameters. The actual
parameters will retain their old values.

7.3. Packages
Packages are PL/SQL constructs that allow related objects to be stored
together.

PACKAGE consists of two parts, namely PACKAGE SPECIFICATION and

PACKAGE BODY. Each of them is stored separately in Data Dictionary.

PACKAGE SPECIFICATION is used to declare functions, procedures that are

part of the package. PACKAGE SPECIFICATION also contains variable and
cursor declarations, which are used by the functions and procedures. Any
object declared in a package specification can be referenced from other
PL/SQL blocks so packages provide global variables to PL/SQL.

PACKAGE BODY contains the function and procedure definitions, which are
declared in the PACKAGE SPECIFICATION. The PACKAGE BODY is optional. If
the package specification does not contain any procedures or functions and
contains only variable and cursor declarations then the body need not be
present

All functions and procedures declared in the package specification are

accessible to all users who have permissions to access the package. Users
cannot access subprograms, which are defined in the package body, but not
declared in the package specification. They can only be accessed by the
subprograms within the package body. This facility is used to hide unwanted
or sensitive information from users.

Page 84 of 91
DBMS/SQL+

A package generally consists of functions and procedures, which are required

by a specific application or a particular module of an application.

Syntax of PACKAGE SPECIFICATION:

CREATE PACKAGE Package_Name AS

Variable_Declaration ;
Cursor_Declaration ;
Function_Declaration ;
Procedure_Declaration ;
END Package_Name ;

where

Function_Declaration is of the form

FUNCTION Func_Name(Param datatype,...)
RETURN datatype1 ;
Procedure_Declaration is of the form
PROCEDURE Proc_Name(Param {IN|OUT|IN OUT}datatype,...);

Syntax of PACKAGE BODY:

CREATE PACKAGE BODY Package_Name AS

Variable_Declaration ;
Cursor_Declaration ;
PROCEDURE Proc_Name(Param {IN|OUT|IN OUT} datatype,...} IS
BEGIN
PL/SQL_Statements ;
END Proc_Name ;
FUNCTION Func_Name(Param datatype,...) IS
BEGIN
PL/SQL_Statements ;
END Func_Name ;
END Package_Name ;

Page 85 of 91
DBMS/SQL+

CREATE OR REPLACE PACKAGE Emp_Actions AS

--Following cursor can be used by other PL/SQl blocks
--Procedures and Functions
PROCEDURE New_Employee
( P_Empno NUMBER,
P_Ename VARCHAR2,
P_Job VARCHAR2,
P_Mgr VARCHAR2,
P_Sal VARCHAR2,
P_Comm VARCHAR2,
P_Deptno VARCHAR2
);
PROCEDURE Fire_Employee( P_Empno NUMBER);
FUNCTION Get_Sal( P_Empno NUMBER ) RETURN NUMBER;
END Emp_Actions ;

CREATE OR REPLACE PACKAGE BODY Emp_Actions AS

PROCEDURE New_Employee
( P_Empno NUMBER,
P_Ename VARCHAR2,
P_Job VARCHAR2,
P_Mgr VARCHAR2,
P_Sal VARCHAR2,
P_Comm VARCHAR2,
P_Deptno VARCHAR2
)
AS
BEGIN
INSERT INTO emp VALUES( P_Empno,P_Ename,
P_Job,P_Mgr,SYSDATE,P_Sal, P_Comm,P_Deptno);
END New_Employee;

PROCEDURE Fire_Employee ( P_Empno NUMBER) IS

BEGIN
DELETE FROM emp WHERE empno=P_Empno ;
END Fire_Employee ;

FUNCTION Get_Sal ( P_Empno NUMBER)

RETURN NUMBER AS
V_Sal NUMBER ;
BEGIN
SELECT sal INTO V_Sal FROM emp WHERE empno =P_Empno;
RETURN V_Sal;
EXCEPTION
WHEN NO_DATA_FOUND THEN
RAISE_APPLICATION_ERROR( -20999 , 'Employee Not Present');
END Get_Sal ;
END Emp_Actions;

Example 7.9

Page 86 of 91
DBMS/SQL+

The procedure and functions calls are the same as in standalone

subprograms. The packaged procedures and functions have to be prefixed
with package names.

BEGIN
Emp_Actions.New_Employee( 10, 'ABCD', …….);
V_Sal := Emp_Actions. Get_Sal( 7566);
END ;

The first time a package is called, it is instantiated. This means that the
package is read from disk into memory and P-CODE is run. At this point
memory is allocated for any variables defined in the package. Each session
will have its own copy of packaged variables, so there is no problem of two
simultaneous sessions accessing the same memory locations.

Page 87 of 91
DBMS/SQL+

8. Table of Examples
Figure 1-1 DBMS Structure ..............................................................3
Figure 1-2 Relational Tables...........................................................7
Example 5.1.........................................................................................62
Figure 5-1 Datatypes ...........................................................................63
Example 5.2 ..........................................................................................67
Example 5.3 ..........................................................................................67
Example 5.4 ..........................................................................................67
Example 5.5 ..........................................................................................67
Example 5.6 ..........................................................................................68
Example 5.7 ..........................................................................................68
Example 5.8 ..........................................................................................68
Example 5.13........................................................................................69
Example 5.14........................................................................................70
Example 5.15......................................................................................71
Example 5.16........................................................................................71
Example 6.1........................................................................................76
Example 6.2........................................................................................77
Example 7.1........................................................................................80
Example 7.2........................................................................................81

Page 88 of 91
DBMS/SQL+

9. Appendix A
Guidelines for Good SQL:

• SQL statements should be reviewed for performance to determine if they are

being executed in the most optimal manner. The following standards are
performance guidelines for SQL statements to ensure that the SQL
statements are performing the most efficient way possible.

• Avoid using arithmetic expressions or functions on the left hand side of the
SELECT criteria (as in WHERE clause) that involves a column name; since it
will not use index.

• '[NOT] EXISTS' should be used rather than '[NOT] IN' when using a subquery
that tests existence or absence of a row.

• When a query is having an 'OR' in a 'WHERE' clause, and both of the

criteria sections of the 'WHERE/OR' clause are specifying different indexes,
and one of the indices is desired over the other, use a function on the
portion of the 'OR' that is not desired to force an index.

• For example, in 'WHERE PART_NO = '3038M83P09' OR CUSTOMER_NO =

'123456'', if both of these force different indices, and it is desired to have
the CUSTOMER_NO index used, then use some function like 'WHERE
NVL(PART_NO,0)='3030M83P09' OR ....'.

• To use an index, the columns in the WHERE clause should be ordered in the
order they appear on the desired index.

• When using more than one table on a FROM clause, the table that is having
the larger number of rows should be listed first.

• When using an '... IN' clause, the elements in the subset (,,,) should be
listed in alphabetical order.

• When poor performance exists with very complex joins, the UNION clause
should be used instead of a normal join, whenever possible.

• For numeric columns that allow nulls, NVL function should be used

Page 89 of 91
DBMS/SQL+

10. Appendix B
Dept Table

Name Type
DEPTNO NUMBER(2)

DNAME VARCHAR2(14)

LOC VARCHAR2(13)

Deptno Dname Loc

10 ACCOUNTING NEW YORK
20 RESEARCH DALLAS
30 SALES CHICAGO
40 OPERATIONS BOSTON

Emp Table

Name Type
EMPNO NUMBER(4)
ENAME VARCHAR2(10)
JOB VARCHAR2(9)
MGR NUMBER(4)
HIREDATE DATE
SAL NUMBER(7,2)
COMM NUMBER(7,2)
DEPTNO NUMBER(2)

Page 90 of 91
DBMS/SQL+

Empno Ename Job Mgr Hiredate Sal Comm Dept

7369 SMITH CLERK 7902 17-Dec-80 800 20
7499 ALLEN SALESMAN 7698 20-Feb-81 1600 300 30
7521 WARD SALESMAN 7698 22-Feb-81 1250 500 30
7566 JONES MANAGER 7839 2-Apr-81 2975 20
7654 MARTIN SALESMAN 7698 28-Sep-81 1250 1400 30
7698 BLAKE MANAGER 7839 1-May-81 2850 30
7782 CLARK MANAGER 7839 9-Jun-81 2450 10
7788 SCOTT ANALYST 7566 9-Dec-82 3000 20
7839 KING PRESIDENT 17-Nov-81 5000 10
7844 TURNER SALESMAN 7698 8-Sep-81 1500 0 30
7876 ADAMS CLERK 7788 12-Jan-83 1100 20
7900 JAMES CLERK 7698 3-Dec-81 950 30
7902 FORD ANALYST 7566 3-Dec-81 3000 20
7934 MILLER CLERK 7782 23-Jan-82 1300 10

Page 91 of 91