Database Systems

Chapter 3: Relational Model Normalization

session 2:
Normal forms

1
 Outline

1 Introduction to Normal Forms

2 The process of normalization

3 1NF, 2NF, 3NF

4 BCNF

1
2
 Normal forms
Normalization works through a series of stages
called normal forms.
 The first normal form (1NF), second normal form
(2NF), and third normal form (3NF).
From a structural point of view, 2NF is better than
1NF, and 3NF is better than 2NF.
For most purposes in business database design, 3NF
is as high as you need to go in the normalization
process.
Highest level of normalization is not always most
desirable
3
The process of normalization

4
 First Normal Form (1NF)
A relation is in first normal form (1NF) if:
No multivalued attributes or no repeating groups
in the table. In other words, each row/column
intersection contains one and only one value, not
a set of values (Every attribute value is atomic)
A primary key has been defined, which uniquely
identifies each row in the relation
All relations are in 1st Normal Form

5
 First Normal Form (1NF)
Example: Invoice tabular
Table with multivalued attributes, not in 1st normal form

Note: this is NOT a relation

6
 First Normal Form (1NF)
Removing repeating groups
=>Invoice table with no multivalued attributes and
unique rows

7
 First Normal Form (1NF)
Select the primary key
There are four determinants in INVOICE, and their
functional dependencies are the following:

The only candidate key for INVOICE is the composite

key consisting of the attributes OrderID and ProductID.
=>The invoice is in 1NF, but not a well-structured one
8
 First Normal Form (1NF)
Anomalies in the table:
Insertion–if new product is ordered for order 1007 of
existing customer, customer data must be re-entered,
causing duplication
Deletion–if we delete the Dining Table from Order 1006, we
lose information concerning this item's finish (Natural Ash)
and price ($800.00).
Update–changing the price of product ID 4 requires update
in several records
Why do these anomalies exist?
Because there are multiple entity types in one
relation. This results in duplication and an
unnecessary dependency between the entities
9
 First Normal Form (1NF)
Ex: Is the supplier table in 1NF?

10
 Second Normal Form (2NF)
A relation is in second normal form (2NF) if:
It is in 1NF
There is no partial dependency.
A partial functional dependency exists when a
non-key attribute is functionally dependent on part
(but not all) of the primary key
If a relation in 1NF has a primary key with single
attribute, it is automatically in 2NF.

11
 Second Normal Form (2NF)
Functional dependency diagram for INVOICE

Primary key for relation is (OrderID,ProductID)

There are following partial dependencies

Therefore, it is NOT in 2nd Normal Form

12
 Second Normal Form (2NF)
To convert a relation with partial dependencies
to 2NF, the following steps are required:
1. Create a new relation for each primary key attribute
(or combination of attributes) that is a determinant in
a partial dependency. That attribute is the primary key
in the new relation.
2. Move the non-key attributes that are dependent on
this primary key attribute (or attributes) from the old
relation to the new relation

13
 Second Normal Form (2NF)
Removing partial dependencies

Partial dependencies are removed, but there

are still transitive dependencies
14
 Exercises
Ex1: Consider a relation R(A,B,C,D) with
F = { AB -> C, A -> D}. Test whether R is in 2NF or not.
=> AB is an only candidate key. Why?
A and B are called prime attributes and C and D are non-
prime attributes.
AB -> C satisfies 2NF. But, A -> D is partial dependency.
So, R is not in 2NF. Therefore, decomposing relation into 2
relations R1(A,B,C) and R2(A, D). R1,R2 is in 2NF.

15
 Exercises
Consider the following relations, check whether them
are in 2NF or not. If not, decompose it into 2NF.
Ex2: The relation R(A,B,C,D) and
F = { AB -> C, B -> D}.
Ex3: The relation R(A,B,C,D,E) and
F = { AB -> C, A -> D, B -> E}.

16
 Third Normal Form (3NF)
A relation is in third normal form (3NF) if:
It is in 2NF
There is no transitive dependency.
A transitive dependency in a relation is a
functional dependency between the primary key
and one or more non-key attributes that are
dependent on the primary key via another non-
key attribute
In the other words, a transitive dependency
exists when a non-key attribute can determine
another non-key attribute
17
 Third Normal Form (3NF)
There are 2 transitive dependencies in CustomerOrder
relation:

Both CustomerName and CustomerAddres are uniquely

identified by CustomerID but CustomerID is not part of the
primary key.

18
 Third Normal Form (3NF)
 Removing transitive dependencies by the following
steps:
1. For each non- key attribute(s) that is a determinant in a
relation, create a new relation. That attribute (s) becomes
the primary key of the new relation.
2. Move all of the attributes that are functionally dependent
on the primary key of the new relation from the old to the
new relation.
3. Primary key in the new relation serve as a foreign key in
the old relation.

19
 Third Normal Form (3NF)
The results of applying the steps to the relation
CustomerOrder

The CustomerOrder relation is divided into 2 relations

Customer and Order =>These relation are in 3NF

20
 Third Normal Form (3NF)
A relational schema showing these four relations and
their associations in ERD

21
 Exercises
Ex1: Is the Inventory relation in 3 NF?

The dependency is shown below:

 Exercises
Ex2: Is the Customer relation in 3 NF?

The dependency is shown below:

 Exercises
Are the following relations in 3NF? If not, decompose
them into 3NF.
Ex3: Consider a relation R(A,B,C,D) with
FDs { AB -> C, C -> D}.
Ex4: Consider the relation R(A,B,C,D, E)
and F = { AB -> C, B ->D, D ->E }.

24
 Boyce-codd Normal Form (BCNF)
A relation is in BCNF if:
It is in 3NF
For every functional dependency X-> Y, X (called
determinant) should be the candidate key of the
relation.
BCNF can be violated only when the table contains
more than one candidate key.
When a table contains only one candidate key, the
3NF and the BCNF are equivalent.

25
 Boyce-codd Normal Form (BCNF)
Example 1: Is the following relation in BCNF?
R(A, B, C, D)
F= { A ->BCD; BC ->AD; D->B}
Above relation is in 3NF. Why?
2 Candidate keys is A and BC. Why?
But FD: D->B, D is not a candidate key.
So the relation R violates BCNF.
Thus, we break the relation R into 2 relation R1(A, D, C)
and R2 (D, B), then R1, R2 are in BCNF.

26
 Boyce-codd Normal Form (BCNF)
Ex2: Is the following relation in BCNF?
R(A, B, C, D)
F= { AB ->CD; AC ->BD; C->B}
Above relation is in 3NF (no partial dependencies and
no transitive dependencies - the C->B: one key/prime
attribute determines another key/prime attribute)
2 Candidate keys is AB and AC
And FD: C->B, C is not a candidate key.
So the relation R violates BCNF. Thus, we break the
relation R into 2 relation R1 and R2, then R1, R2 are in
BCNF with R1(A, C, D), R2 (C,B)
27
 Boyce-codd Normal Form (BCNF)
Ex3: Is the Student relation in BCNF?

 With FDs:

{Stu_ID, Staff_ID} -> {Class_code, Enroll_Grade}

Class_Code ->Staff_ID
28
 Boyce-codd Normal Form (BCNF)
=> 2 problems with the Student relation:
Update anomaly: if a different staff member is
assigned to teach class 32456, two rows will require
updates
deletion anomaly: if student 135 drops class 28458,
information about who taught that class is lost
The relation has 2 candidate keys: {Stu_ID, Staff_ID} and
{StudentID, Class_Code}
And Class_Code ->Staff_ID violates BCNF.
So, decomposition to BCNF into 2 relations with
Student(Stu_ID, Class_Code, Enroll_Grade} and
Class(Class_Code, Staff_ID}
29
 BCNF (Boyce-Codd Normal Form)
Ex4: Consider the relation R(A,B,C,D) and
F = { AB -> C, C -> D}. Is the relation in BCNF?
=>AB is only one candidate key. Why?
A,B are key/prime attributes and C,D are non-key/non-
prime attributes.
C -> D is transitive dependency (a non-key attribute
determine another non-key attribute). So, R is not in 3NF.
Therefore, R is not in BCNF as well.
To convert to 3NF, decompose R into two relations
R1(A,B,C) and R2(C,D). Now the relations will be in BCNF
as well.
 Exercises
Determine whether the following relations is in BCNF
or not? If not, decompose the relation with its NF.
Ex1: Determine whether relation R(A,B,C) with
F = { AB -> C, C -> A}
Ex2: The relation StudMajor( StdNo, Major, Advisor)
with FDs {StdNo, Major -> Advisor, Advisor -> Major }.

31
32