Normalization

Background Knowledge

- Well Structured Relation

A relation is said to be a well structured relation if it doesn’t have any anomaly.

o Anomaly
It is an inconsistency that may arise when a user tries to insert data (insertion
anomaly) or tries to modify data (modification / updation anomaly) or tries to
delete data (deletion anomaly).
Student
R# Name C-Code C-Title

1 Faisal Cs-241 DBMS

1 Faisal Cs-342 OS

2 Imran Cs-342 OS

4 Ahsan Cs-312 DLD

5 Ali

 Insertion Anomaly
We can’t insert data of a student unless he registers some course, or vice
versa.

 Modification / Updation Anomaly

If we want to change the name of Faisa, R#-1, from Faisal to Muhammad
Faisal, then we have to make this change for multiple times, as it occurs
more than one times in the Student Data Base.

 Deletion Anomaly
If we want to delete student data R#-4, Ahsan, we also lose data about
the course Cs-312, DLD.
DBMS

www.bscs15.tk www.leemedia.tk www.ttdate.tk

 Normalization
It is a step by step process in which we convert a complex data structure (relation with
anomalies) into a well structured relation (Relation without anomalies).

- Normal Form
It is the state or form of a relation that arises when we apply some rules of
normalization on that relation, for example First Normal Form or 1NF, 2NF etc.

o 1NF
A relation is said to be in 1NF if it doesn’t have any repeating group.

Problem
R# Name C-Code C-Title

1 Faisal Cs-241 DBMS

1 Faisal Cs-342 OS

2 Imran Cs-342 OS

4 Ahsan Cs-312 DLD

This table contains repeating group, so it’s not in 1NF

Solution
We make a composite primary key by combining some attributes from the
repeating group, to remove the repeating group.
R# Name C-Code C-Title

1 Faisal Cs-241 DBMS

1 Faisal Cs-342 OS

2 Imran Cs-342 OS

4 Ahsan Cs-312 DLD

Now it’s in 1NF.

DBMS

www.bscs15.tk www.leemedia.tk www.ttdate.tk

 o 2NF
A relation is said to be 2NF if it is in 1NF, and there is no partial functional
dependency in it.

 Functional Dependency
If we have a relation
R (A1, A2, A3, A4, A5)
Student (R#, name, C-Code, C-Title)

If A1 determines A2, A3, A4, A5 then we say that A2, A3, A4, A5 are
functionally dependent upon A1.

 Partial Functional Dependency

If we have a relation;
R (A1, A2, A3, A4, A5)

R# Name C-Code C-Title Grade

1 Faisal Cs-241 DBMS A

1 Faisal Cs-342 OS A

2 Imran Cs-342 OS A

4 Ahsan Cs-312 DLD A

If part of a composite primary key determines some non-key attributes

then this phenomenon is known as partial functional dependency.

A1  A2 partially functional dependent

A3  A4 partially functional dependent
 A5 Fully functional dependent
DBMS

www.bscs15.tk www.leemedia.tk www.ttdate.tk

 Solution

R# Name R# C-Code Grade C-Code C-Title

1 Faisal 1 Cs-241 A Cs-241 DBMS

1 Faisal 1 Cs-342 A Cs-342 OS

2 Imran 2 Cs-342 A Cs-342 OS

4 Ahsan 4 Cs-312 A Cs-312 DLD

DBMS

www.bscs15.tk www.leemedia.tk www.ttdate.tk

 o 3NF
A relation is said to be in 3NF if it is in 2NF and there is no transitive dependency
in it.
 Transitive Dependency
If a non-key attribute determines another non-key attribute, then this
phenomenon is known as transitive dependency.

For example
R (A, B, C, D, E)
And in this relation AB, C, D, E but CE as well.

C-no C-name SP-ID SP-Name Region

C1 Waleed SP-3 Ali G-10

C2 Waleed SP-4 Ahmed G-10

C3 Ali SP-7 Naeem G-8

C4 Ahsan SP-3 Ali G-10

C5 Yahya SP-9 Ali I-9

C6 Waleed Sp-13 Ahmed F-10

C-no C-name, SP-ID, SP-Name, Region

 Hints
1. Customer can be served by only one sales person, while a sales
person can serve many customers.
2. One sales person belongs to only one region, while a region can
have many sales persons.
DBMS

www.bscs15.tk www.leemedia.tk www.ttdate.tk

 Solution

C-no C-name SP-ID SP-ID SP-Name Region

C1 Waleed SP-3 SP-3 Ali G-10

C2 Waleed SP-4 SP-4 Ahmed G-10

C3 Ali SP-7 SP-7 Naeem G-8

C4 Ahsan SP-3 SP-3 Ali G-10

C5 Yahya SP-9 SP-9 Ali I-9

C6 Waleed Sp-13 Sp-13 Ahmed F-10

DBMS

www.bscs15.tk www.leemedia.tk www.ttdate.tk