1. Database Normalisation
Databases
Database Normalisation

2. Learning Objectives
Design simple relational databases to the third normal form (3NF).

3. Database Normalization
A technique for designing relational database tables to minimize duplication of information which can lead to inconsistencies, leading to a loss of data integrity.

4. Example of Database Normalisation

5. Example Database
A database is designed to store data about students at a college and the subjects that they study.
All students are based in a tutor group
A tutor supervises all the students in their tutor group, can only be a tutor of one form and a form has only one tutor
Each subject has one subject teacher only
Students study a number of subjects

6. Un-Normalised Form (UNF)
A relation with repeating groups (fields which are duplicated an unknown number of times) which have their own primary keys.
(i.e. many to many relationships)
The next table is in un-normalised form.

7. Table: StudentSubject
StudentID
Student
Name
Tutor
SubjectID
Subject
Level
Teacher
Grade
0001
Tom
SAN
PHYA2
CHEA2
GENAS
Physics
Chemistry
Gen. Studies A AS
MEB
DIL B E
0002
Joe
GEOAS
FREAS
Geography
French
ROG
HEN C
0003
Samir
COMA2
MATA2
GENA2
Computing
Maths
VAR
COR D A*
Repeating groups
Un-Normalised Form (UNF):
The table has rows/records/tuples which contain a repeated group of attributes (which have their own primary keys, in this case -SubjectID) that represent the subjects each student studies.
Underlined field is the primary key.

8. UNF Table description / shorthand notation / design:
Table: StudentSubjects (StudentID, StudentName, TutorGroup,
Tutor, (SubjectID, Subject, Level,SubjectTeacher, Grade))
The words in brackets separated by a , = Fields / Columns / Attributes
Underlined Field = Primary Key
Inner brackets = Repeating group (many to many relationships)
Some websites / books also underline the primary key of the repeating group e.g. the SubjectID in this case. I will not here as it is I believe it is confusing when thinking about 1NF – see next slides.
Subject
Student
studied by
study

9. First Normal Form (1NF)
Remove repeating groups (and their primary keys).
Expand the primary key to include the unique key of the repeating group (concatenated / combination / compound primary key).

10. Table: StudentSubject
StudentID
SubjectID
Student
Name
Tutor
Group
Subject
Level
Teacher
Grade
0001
PHYA2
Tom 6
SAN
Physics A
CHEA2
Chemistry
MEB B
GENAS
Gen. Studies AS
DIL E
0002
GEOAS
Joe 7
Geography
ROG C
FREAS
French
HEN
0003
COMA2
Samir
Computing
VAR D
MATA2
Maths
COR
GENA2 A*
First Normal Form (1NF):
Repeating groups have been removed and the primary key has now been expanded to include SubjectID (StudentID + SubjectID = StudentID/SubjectID), therefore the primary key no longer repeats.
I have placed these 2 fields side by side but this is not necessary, I have only done this to make it easier to see the new primary key.

11. 1NF Table description / shorthand notation / design:
Table: StudentSubject (StudentID, StudentName, TutorGroup,
Tutor, SubjectID, Subjects, Level, SubjectTeacher, Grade)
Primary key = StudentID + SubjectID = StudentID/SubjectID
Note the fields used for the combined primary key do not have to be shown together, just underlined. This is why I did not underline before in UNF but even if you did, the inner brackets show the group repetition in UNF and they are no longer in the notation above for 1NF.

12. Table: Student Table: Subject StudentID/ SubjectID StudentID/
You can also move the repeating group attributes into a new table at this stage.
StudentID
Student
Name
Tutor
Group
0001
Tom 6
SAN
0002
Joe 7
MEB
0003
Samir
Table: Student (StudentID, StudentName,
TutorGroup, Tutor)
Table: Subject (StudentID, SubjectID, Subject,
Level, SubjectTeacher, Grade)
Dashed underline indicates a Foreign key (a non-primary key in one table which is the primary key of another table).
Table: Subject
StudentID
SubjectID
Subject
Level
Teacher
Grade
0001
PHYA2
Physics A
SAN
CHEA2
Chemistry
MEB B
GENAS
Gen. Studies AS
DIL E
0002
GEOAS
Geography
ROG C
FREAS
French
HEN
0003
COMA2
Computing
VAR D
MATA2
Maths
COR
GENA2 A*
There is a one-to-many relationship with Student being the ‘one side’ table and Subject being the ‘many side‘ table.
The primary key StudentID in the Student table links to the foreign key StudentID in the Subject table.
studies
studies
StudentID/
SubjectID
StudentID/
SubjectID
Subject
Student
studied by
studied by

13. 2NF Definition
A relation that is in 1NF and every non-primary key attribute is fully dependent on the primary key (no partial dependencies are allowed).
Do any of the fields in the table depend on only a portion of the primary key?
Yes = Not 2NF
No = 2NF

14. Not in 2NF because: StudentID SubjectID Student Name Tutor Group
Subject, Level, SubjectTeacher depend on the SubjectID but not on the StudentID, and, StudentName, TutorGroup, Tutor depend on the StudentID but not on the SubjectID.
Only the Grade depends on the entire concatenated / combination primary key (StudentID + SubjectID).
StudentID
SubjectID
Student
Name
Tutor
Group
Subject
Level
Teacher
Grade
0001
PHYA2
Tom 6
SAN
Physics A
CHEA2
Chemistry
MEB B
GENAS
Gen. Studies AS
DIL E
0002
GEOAS
Joe 7
Geography
ROG C
FREAS
French
HEN
0003
COMA2
Samir
Computing
VAR D
MATA2
Maths
COR
GENA2 A*

15. To convert a table to 2NF, you must:
Duplicate each portion of the concatenated / combination primary key into another table and move any fields that depend wholly on the duplicated portion into this table. Leaving the concatenated / combination primary key and any fields that wholly depend on it, in the original table; also identify each portion as a foreign key.
In this case:
SubjectID, Subject, Level & SubjectTeacher.
StudentID, StudentName, TutorGroup & Tutor.
duplicated
moved
StudentID
SubjectID
Student
Name
Tutor
Group
Subject
Level
Teacher
Grade
0001
PHYA2
Tom 6
SAN
Physics A
CHEA2
Chemistry
MEB B
GENAS
Gen. Studies AS
DIL E
0002
GEOAS
Joe 7
Geography
ROG C
FREAS
French
HEN
0003
COMA2
Samir
Computing
VAR D
MATA2
Maths
COR
GENA2 A*

16. 2NF Table: Grade Table: Student Table: Subject
StudentID
SubjectID
Grade
0001
PHYA2 A
CHEA2 B
GENAS E
0002
GEOAS C
FREAS
0003
COMA2 D
MATA2
GENA2 A*
StudentID
Student
Name
Tutor
Group
0001
Tom 6
SAN
0002
Joe 7
MEB
0003
Samir
Dashed underline indicates a Foreign key (a non-primary key in one table which is the primary key of another table).
All fields are fully dependent on the entire primary key of their respective tables.
Table: Subject
SubjectID
Subject
Level
Teacher
PHYA2
Physics A
SAN
CHEA2
Chemistry
MEB
GENAS
Gen. Studies AS
DIL
GEOAS
Geography
ROG
FREAS
French
HEN
COMA2
Computing
VAR
CHEAS
MATA2
Maths
COR
GENA2
There is a one-to-many relationship with Student being the ‘one side’ table and Grade being the ‘many side‘ table.
The primary key StudentID in the Student table links to the foreign key StudentID in the Grade table (similarly for SubjectID in the Student & Grade tables).
given
awarded
achieves
SubjectID/
Grade
StudentID/
Grade
Subject
Grade
Student
shows
achieved by

17. SubjectID/ Grade Subject Grade Subject Grade Entity / Table Link
given
SubjectID/
Grade
Subject
Grade
shows
Entity / Table
Link
Entity / Table
Note that SubjectID/Grade is not a table/entity, it is just a way of showing how the tables/entities Subject and Grade are linked.
It is also possible to show only the tables and assume there is a link (the primary key - SubjectID - from Subject is duplicated in Grade and is known as the foreign key in Grade).
Due to the primary key of Grade now being StudentID/SubjectID, there are now many SubjectIDs in Subject to one StudentID/SubjectID in Grade.
The table without a combination primary key forms (in this case Subject) the many side and the table with a combination primary key (in this case Student) forms the one side.
given
Subject
Grade
shows

18. StudentID/ Grade Grade Student Grade Student Entity / Table
given
StudentID/
Grade
Grade
Student
shows
Entity / Table
Entity / Table
Link
Note that StudentID/Grade is not a table/entity, it is just a way of showing how the tables/entities Grade and Student are linked.
It is also possible to show only the tables and assume there is a link (the primary key - StudentID - from Student is duplicated in Grade and is known as the foreign key in Grade).
Due to the primary key of Grade now being StudentID/SubjectID, there are now many StudentIDs in Student to one StudentID/SubjectID in Grade.
The table with a combination primary key (in this case Grade) forms the one side and the table without a combination primary key forms (in this case Student) the many side.
given
Grade
Student
shows

19. 2NF Table description / shorthand notation / design:
Table: Student (StudentID, StudentName, TutorGroup, Tutor)
Table: Subject (SubjectID, Subject, Level, SubjectTeacher)
Table: Grade (SubjectID, StudentID, Grade)
given
awarded
achieves
SubjectID/
Grade
StudentID/
Grade
Subject
Grade
Student
shows
achieved by

20. 3NF Definition
A relation that is in 1NF and 2NF, and if no non-key field is dependent on another non-key field
(i.e. no one to one relationships).

21. Table: Student Student TutorGroup Tutor
Table Student is not in 3NF because:
TutorGroup is dependent on Tutor and Tutor is dependent on TutorGroup.
Table: Student
StudentID
Student
Name
Tutor
Group
0001
Tom 6
SAN
0002
Joe 7
MEB
0003
Samir
Student
TutorGroup
Tutor
To convert the table to 3NF, you must:
Duplicate one of the non-key fields that depends on another non-key field into another table and move all other non-key fields that depend on another non-key field into this table (in this case: duplicate TutorGroup & move Tutor).
Use the duplicated field as the primary key of the new table and as a foreign key in the original table.

22. Tables: Subject & Grade were already in 3NF.
Table: Student
3NF
Dashed underline indicates a Foreign key (a non-primary key in one table which is the primary key of another table).
StudentID
Student
Name
Tutor
Group
0001
Tom 6
0002
Joe 7
0003
Samir
Table: Grade
StudentID
SubjectID
Grade
0001
PHYA2 A
CHEA2 B
GENAS E
0002
GEOAS C
FREAS
0003
COMA2 D
MATA2
GENA2 A*
Table: Subject
SubjectID
Subject
Level
Teacher
PHYA2
Physics A
SAN
CHEA2
Chemistry
MEB
GENAS
Gen. Studies AS
DIL
GEOAS
Geography
ROG
FREAS
French
HEN
COMA2
Computing
VAR
CHEAS
MATA2
Maths
COR
GENA2
Table: Tutor
Tutor
Group 6
SAN 7
MEB
Tables: Subject & Grade were already in 3NF.
Third Normal Form (3NF):
Each table relates to only 1 entity, has only 1 primary key and there are no repeating non-key fields.

23. 3NF Table description / shorthand notation / design:
Table: Student (StudentID, StudentName, TutorGroup)
Dashed underline indicates a Foreign key (a non-primary key in one table which is the primary key of another table).
Table: Tutor (TutorGroup, Tutor)
Table: Subject (SubjectID, Subject, Level, SubjectTeacher)
Table: Grade (SubjectID, StudentID, Grade)
Student
Tutor_TutorGroup

24. Normalisation Summary
UNF
The table has rows which contain repeated groups of attributes (which have their own primary keys).
1NF
Remove repeating groups (and their primary keys).
Expand the primary key to include the unique key of the repeating group (concatenated / combination primary key).
2NF
Duplicate each portion of the concatenated / combination primary key into another table and move any fields that depend wholly on the duplicated portion into this table. Leaving the concatenated / combination primary key and any fields that wholly depend on it, in the original table; also identify each portion as a foreign key.
3NF
Duplicate one of the non-key fields that depends on another non-key field into another table and move all other non-key fields that depend on another non-key field into this table (in this case: duplicate TutorGroup & move Tutor).
Use the duplicated as the primary key of the new table and as a foreign key in the original table.

25. Plenary
When a customer orders flowers, an order form has to be completed.
The order form is shown on the next slide.

26. CUSTOMER ORDER
Order Number:
Date: / /
Customer Number:
QUANTITY
FLOWER ID

27. Plenary
Create a table in UNF, called ORDER, which contains all the attributes shown on the order form.
Explain why it is not normalised.

28. Plenary
ORDER (OrderNumber, CustomerNumber, Date, (Quantity), (FlowerID))
ORDER is not normalised as it is has repeating groups.

29. Plenary
Show this data model in 1NF, 2NF and finally 3NF.

30. Plenary Remove the repeating groups to obtain 1NF.
ORDER (OrderNumber, CustomerNumber, Date, Quantity, FlowerID)

31. Plenary
CustomerNumber and Date are not dependent on FlowerID so remove them to another table to obtain 2NF.
ORDER (CustomerNumber, Date)
FLOWERORDER (OrderNumber, FlowerID, Quantity)
This is also in 3NF.

32. Plenary
The owner of a flower shop uses a relational database to store information about orders placed by customers, and the types of flower in stock.
One entity is defined as CUSTOMERS.
List the attributes which you identify as belonging to this entity.

33. Plenary Forename, Surname Address1, Address2, County Postcode
Date of last order
Credit limit