1. Normalization (Database Design)
Chapter 3
Normalization (Database Design)

2. Chapter 3 : Objectives
What normalization is and what role it plays in database design
About the normal forms 1NF, 2NF, 3NF, BCNF, and 4NF
How normal forms can be transformed from lower normal forms to higher normal forms
That normalization and E-R modeling are used concurrently to produce a good database design
That some situations require denormalization to generate information efficiently

3. What is Normalization ?
A technique for producing a set of relations with desirable properties (minimum data redundancy), given the data requirements of an enterprise.
First developed by E.F. Codd (1972)

4. Why Normalization ?
Main objective in developing a logical data model for relational database systems is to create an accurate representation of the data, its relationships, and constraints.
To achieve this objective, must identify a suitable set of relations.
A relation can be normalized to a specific form to prevent possible occurrence of update anomalies.

5. Why Normalization ?
As normalization proceeds, relations become progressively more restricted (stronger) in format and also less vulnerable to update anomalies.

6. Why Normalize ?
An ER Model is not always available as a starting point for design
To reduce redundant data in existing design
To increase integrity of data and stability of design
To identify missing tables, columns and constraints

7. Why Normalization ? SID Name Grade Course# Text Major Dept
S1 Joseph A CIS800 b1 CIS CIS
S1 Joseph B CIS820 b2 CIS CIS
S1 Joseph A CIS872 b5 CIS CIS
S2 Alice A CIS800 b1 CS MCS
S2 Alice A CIS872 b5 CS MCS
S3 Tom B CIS800 b1 Acct Acct
S3 Tom B CIS872 b5 Acct Acct
S3 Tom A CIS860 b1 Acct Acct
Is there any redundant data?
Can we insert a new Course# with a new textbook?
What should be done if 'CIS' is changed to 'MIS'?
What would happen if we remove all CIS800 students?

8. Update Anomalies A poorly designed relation contains redundant data.
Data redundancy causes update anomalies :
Insertion Anomalies -An independent piece of information cannot be recorded in a relation unless an irrelevant information must be inserted together at the same time.
Modification Anomalies - The update of a piece of information must occur at multiple locations (if not, it leads to inconsistency).
Deletion Anomalies - The deletion of a piece of information unintentionally removes other information.
Normalization is used to remove update anomalies by decomposing the relations into smaller relations based on the constraints.

9. First Normal Form All key attributes defined
No repeating groups in table
All attributes dependent on
primary key

10. Example : 1NF Example : FIRST(S#,STATUS,CITY,P#,QTY)
- Suppliers supply certain quantities of parts and are located in some city.
- Status of the supplier is determined by the location (city) of that supplier.
Fig : Relation FIRST
S# STATUS CITY P# QTY
S1 20 London P1 300
S1 20 London P2 200
S1 20 London P3 400
S1 20 London P4 200
S1 20 London P5 100
S1 20 London P6 100
S2 10 Paris P1 300
S2 10 Paris P2 400
S3 10 Paris P2 200
S4 20 London P2 200
S4 20 London P4 300
S4 20 London P5 400

11. Example : 1NF S# STATUS CITY P# QTY S1 20 London P1 300
S2 10 Paris P1 300
S2 10 Paris P2 400
S3 10 Paris P2 200
S4 20 London P2 200
S4 20 London P4 300
S4 20 London P5 400
What is the Primary key of FIRST ?
Is FIRST in 1NF ?
Does FIRST have update anomalies ?

12. Example : 1NF
To overcome the above anomalies in the relation FIRST , FIRST has to be decomposed into smaller relations (projections).
=> Decomposition

13. Non-loss Decomposition
To overcome the update anomalies in a relation R, R is decomposed into smaller relations (projections).
A bad decomposition loses information.
In a good decomposition
The join of decomposed relations restores the original relation.
Decomposed relations can be maintained independently.

14. Non-loss Decomposition
Rissanen's Rules:
Two projections R1 and R2 of a relation R are independent iff
Every FD in R can be logically derived from those in R1 and R2 and
The common attributes of R1 and R2 form a candidate key for at least one of the pair.

15. Functional Dependency
Given two attributes X and Y of a relation R, Y is functionally dependent on X iff each X-value in R has associated with it exactly one Y-value in R(if two tuples agree on their X-value, they also agree on their Y-value).
X  Y

16. Functional Dependency
Example: List all FDs in FIRST
FD Diagram of FIRST: S# P#
QTY
CITY
STATUS

17. Properties of FDs X is called the determinant of Y
X may or may not be the key attribute of R
X and Y may be composite
X and Y could be mutually dependent on each other
Husband  Wife Wife  Husband
The same Y-value may occur in multiple tuples

18. Properties of FDs An FD could be trivial or non-trivial
An FD is trivial
if it is impossible for it not to be satisfied
if the RHS is a subset of the LHS
In Normalization, the FDs of interest are :
non-trivial
hold for all times

19. Full Functional Dependency :
Given attributes X and Y of a relation R, Y is fully functionally dependent on X if Y is functionally dependent on X but not on any proper subset of X.

20. Transitive Functional Dependency
Given attributes X, Y and Z of a relation R, Z is transitively dependent on X iff
X  Y , Y  Z and X  Z

21. UNNORMALIZED FORM (UNF)
A table that contains one or more repeating groups.
Examples :
Is FIRST Unnormalized ?
Is relation SCH given below Unnormalized ?
Student Courses Hobbies
Jin DBMS Tennis
OS Cooking
Habib C++ Reading
Networking Surfing CA

22. UNF to 1NF
Remove repeating group by entering appropriate data into the empty columns of rows containing repeating data (‘flattening’ the table).
Student Courses Hobbies
Jin DBMS Tennis
Jin DBMS Cooking
Jin OS Tennis
Jin OS Cooking
Habib C++ Reading
Habib C++ Surfing
Habib Networking Reading
Habib Networking Surfing
Habib CA Reading
Habib CA Surfing

23. FIRST NORMAL FORM (1NF)- revisited
A relation R is in 1NF iff all attribute domains contain atomic values only.
Example :
FIRST is in 1NF
FIRST has update anomalies

24. FIRST NORMAL FORM (1NF)
FD diagram of FIRST: S# P#
QTY
CITY
STATUS

25. SECOND NORMAL FORM (2NF)
A relation R is in 2NF iff it is in 1NF and every non-prime-key attribute is fully functionally dependent on the primary key.
A non-prime-key attribute does not belong to any candidate key.
Is FIRST in 2NF ?
How to normalize FIRST into 2NF ?
Remove all partial (or non-full) dependencies from FIRST by decomposing FIRST into two smaller relations SP and SECOND

26. SECOND NORMAL FORM (2NF)
Is the decomposition of First into SP and SECOND good ?
SP(S#,P#,QTY) SECOND(S#,CITY,STATUS)

27. SECOND NORMAL FORM (2NF)
FD Diagram of SP and SECOND
SECOND is in 2NF
SECOND has update anomalies S# P#
QTY
CITY
STATUS

28. THIRD NORMAL FORM(3NF)
A relation is in 3NF iff it is in 2NF and in which every non-prime-key attribute is non-transitively dependent on the primary key.
Is SECOND(S#,CITY,STATUS) in 3NF ?

29. THIRD NORMAL FORM(3NF) How to normalize SECOND into 3NF ?
Remove all transitive dependencies from SECOND by decomposing SECOND into two smaller relations SC and CS.
SC(S#,CITY) CS(CITY,STATUS)
Make sure that the decomposition is good !

31. THIRD NORMAL FORM(3NF)
FD Diagram of SC and CS S#
CITY
STATUS

32. Help me Codd !
The key(1NF), the whole key(2NF), and nothing but the key (3NF) - so help me Codd !!

33. THIRD NORMAL FORM(3NF) Relation R(A,B,C,D) given below is in 3NF
Still has Update Anomalies !!

34. Example : Example: TEACH (Student, Course, Instructor)
A student can take several courses and can be taught by several instructors.
For each course, each student of that course is taught by only one teacher.
An instructor teaches only one course.
Two Candidate keys : (Student, Course) (Student, Instructor)

35. Example : Choose (Student, Course) as the Primary key
FD Diagram for TEACHES
Is TEACHES in 3NF ?
Does TEACHES have update anomalies ?
What happens if (Student, Instructor) is chosen as the Primary key!
STUDENT
COURSE
INSTRUCTOR

36. BOYCE/CODD NORMAL FORM
A relation is in BCNF iff every determinant is a Candidate key.
Update Anomalies occur in a 3NF relation R if
R has multiple candidate keys
Those candidate keys are composite and
The candidate keys are overlapped
How to Normalize TEACHES in BCNF ?
Decompose TEACHES into
SI (Student, Instructor) and IC (Instructor, Course)

37. Denormalization Normalization is one of many database design goals
But Is it always the best possible design ?

38. Why Denormalize ?
Normalization is the process of putting one fact in one appropriate place. This optimizes updates at the expense of retrievals.
When a fact is stored in only one place, retrieving many different but related facts usually requires going to many different places. This tends to slow the retrieval process.
Updating is quicker, however, because the fact you're updating exists in only one place.

39. Why Denormalize ?
A relational normalized database imposes a heavy access load over physical storage of data even if it is well tuned for high performance.
A normalized design will often store different but related pieces of information in separate logical tables (called relations).
If these relations are stored physically as separate disk files, completing a database query that draws information from several relations (a join operation) can be slow.

40. Denormalization Denormalization
is the process of attempting to optimize the performance of a database by adding redundant data or by grouping data.
technique to move from higher to lower normal forms of database modeling in order to speed up database access.

41. An Example of Denormalization
Consider a relation :
Contact (Name, Street, Zip, City, Province)
PK of Contact ?
Is Contact in 3NF ?
Name
Street
Zip
City
Province
John
401 Sunset Av.
N91 Q23
Windsor ON
Harry
402 Sunset Av.
N4T 3R5
Bill
123 First St.
N4Y 7Y8
London

42. Example of Denormalization(contd)
Was it worth the decomposition ?