Lecture 8 Normalization

8/20/2019 Lecture 8 Normalization

1/17


2/17

Lecture 7

Systems Analysis & Design 2

3

Basic Conversion Rules For 1:1 cardinality: all attributes are

merged into a single table.

For 1:N cardinality: post the “identifier”from the “one” side as an attribute tothe “many” side.

For N:M cardinality: create a new table

and post the primary keys from eachentity as attributes in the new table.

4

Null Values

Should be avoided.

Make it difficult to construct & interpret

relational data base queries.


3/17

Lecture 7


5

Optional-Max Conversion Rules

There are 9 rules

On completion of these 9 rules a relationaldatabase would be created.

6

Optional-Max Conversion Rules:

Rule 1

Construct an ER

diagram and label theupper and lower boundsfor relationship degree.

All multi-valuedattributes should berepresented as entities.

FacultyFID [PK]F_NameL_name

CourseCID [PK]C_NameCredits

StudentsSID [PK]F_NameL_name

Home Address

H_StreetH_CityH_State

Locker

LID [PK]Building

Teach

Enroll

Stud_Maj

Has Assigned

Major

Maj_ID [PK]Description


4/17

Lecture 7


7


Rule 2

In every instance where the lower

bound for relationship degree is 0

and the upper bound is 1,

temporarily label the upper bound

as N.

If the upper bound is greater than 1

(for example, 3, or 10 or N), leave it

as is.



StudentsSID [PK]F_Name

L_name

Home AddressH_StreetH_CityH_State

LockerLID [PK]Building

Teach

Enroll

Stud_Maj

Has Assigned

MajorMaj_ID [PK]Description

N

N

N

8


Rule 3

Faculty

Course

Students

Home_Address

Major

Locker






Teach

Enroll

Stud_Maj

Has Assigned

MajorMaj_ID [PK]

Description

N

N

N

Use the entity namefor each entity as thetable name


5/17

Lecture 7


9


Rule 4

If an identifier or primary key has

not been selected then select an

identifier for each table.Faculty (FID)

Course (CID)

Students (SID)

Home_Address (H_Street, H_City, H_State)

Major (Maj_ID)

Locker (LID)






Teach

Enroll

Stud_Maj

Has Assigned


N

N

N

10


Rule 5

Take all the attributes that

describe an entity & post themas column names in the

relational tables.

Faculty (FID, F_Name, L_Name)

Course (CID, C_Name, Credits)

Students (SID, F_Name, L_Name)

Home_Address (H_Street, H_City,

H_State)

Major (Maj_ID, Description)

Locker (LID, Building)






Teach

Enroll

Stud_Maj

Has Assigned

MajorMaj_ID [PK]

Description

N

N

N


6/17


7/17

Lecture 7


13


Rule 8

For N:M relationships, create a newtable using the relationship name

as the name for the table.


Teach (FID, CID)


Enroll (CID, SID,)

Students (SID, Maj_ID F_Name,L_Name, H_Street, H_City, H_State)

Major (Maj_ID, Description)

Assigned (SID, LID)


concatenated

key



Enroll

StudentsSID [PK]

F_NameL_name




Teach

C_Enroll

S_Enroll

Stud_Maj

Has Assigned

N

N

N

intersection

table

14

Optional-Max Conversion Rules: Rule 9Take any additional attributes and post them as column names in the appropriate

relational table. If there are any composite attributes in the relational table

definition, decompose them into simple attributes.


Teach (FID, CID)


Enroll (CID, SID, Grade)

Students (SID, Maj_ID F_Name,L_Name, H_Street, H_City, H_State)

Major (Maj _ID, Description)

Assigned (SID, LID, Building)








Teach

Stud_Maj

Has Assigned

N

N

N

Enroll

C_Enroll

S_Enroll


8/17

Lecture 7


15

Optional-Max Conversion Rules

SummaryOptional-Max Conversion Rules

Rule 1 Construct ERD

Rule 2 Change 0:1 to 0:N

Rule 3 Use entity name as Table name

Rule 4 Select and write an identifier for each table

Rule 5 Write down attributes as column names

Rule 6 Merge 1:1 relationships

Rule 7 For 1:N relationships place key from “one”side to the “many” side

Rule 8 For N:M create a new table usingrelationship name

Rule 9 If additional attribute exist, place inappropriate table

16

Normalization

Normalization is a technique for producing a set of

relations with desirable properties, given the datarequirements of an enterprise.

The process of normalization is a formal method that

identifies relations based on their primary or candidate

keys and the functional dependencies among their

attributes.


9/17

Lecture 7


17

Normal forms The 4 most commonly used normal

forms are:

First normal form (1NF)

Second normal form (2NF)

Third normal form (3NF)

Boyce-Codd normal form (BCNF)

18

Normal forms

Unnormalized relation

Remove repeating

groups

Remove partial

dependencies

Remove transitive

dependencies

1NF

2NF

3NF


10/17

Lecture 7


19

WHNo WHLoc ProductNo ProductDesc Price PCode CDesc Qty

10 Ilford 1001 Hammer 10.99 H Hardware 06

1002 Screwdriver 03.99 H Hardware 08

1010 Wood Drill 30.49 ET Electrical 10

30 Exeter 1001 Hammer 10.99 H Hardware 01

4001 Stepladder 49.99 H Hardware 02

Unnormalized form

20

First normal form ( 1NF )

First normal form requires that repeating groups must beremoved

(WHNo, WHLoc, (ProductNo, PDesc, Price, PCode,CDesc, Qty)).

WHNo WHLoc ProductNo ProductDesc Price PCode CDesc Qty

10 Ilford 1001 Hammer 10.99 H Hardware 06

1002 Screwdriver 03.99 H Hardware 08

1010 Wood Drill 30.49 ET Electrical 10

30 Exeter 1001 Hammer 10.99 H Hardware 01

4001 Stepladder 49.99 H Hardware 02

Repeating Group


11/17

Lecture 7


21

First normal form ( 1NF ) WAREHOUSE(WHNo, WHLoc).

STOCK(WHNo, ProductNo, ProductDesc, Price, PCode,CDesc, Qty).

The duplication of WHNo in both serves to establish therelationship, or link, between the two entity types.

Neither WHNo nor ProductNo alone would be sufficient touniquely identify occurrences of STOCK.

It would be possible to reconstitute the original i.e. identifyall the occurrences of STOCK that are related to any givenoccurrence of WAREHOUSE

22


WHNo WHLoc

10 Ilford

30 Exeter

Warehouse

WHNo ProductNo ProductDesc Price PCode CDesc Qty

10 1001 Hammer 10.99 H Hardware 06

10 1002 Screwdriver 03.99 H Hardware 08

10 1010 Wood Drill 30.49 ET Electrical 10


30 4001 Stepladder 49.99 H Hardware 02

Stock


12/17

Lecture 7


23


WHNo WHLoc

10 Ilford

30 Exeter

Warehouse







Stock

Insertion

To insert a new product intothe entity STOCK,Need to know what warehouseit is supplied from.

24


WHNo WHLoc

10 Ilford

30 Exeter

Warehouse







Stock

Deletion

If warehouse 30 is deleted,all the details of Product 4001will also be deleted


13/17

Lecture 7


25


WHNo WHLoc

10 Ilford

30 Exeter

Warehouse







Stock

Editing

same price will be charged fora product regardless of whereit is supplied from.If price of hammers atWHNo10 = 11.99, the price atWHNo30 will still be = 10.99.This will make the datainconsistent.

26

Clearly, WAREHOUSE is already in 2NF

(Why? Because all attributes fully depend on theentire key, WHNo; none of the attributes candepend on part of a WHNo).

WHNo WHLoc

10 Ilford

30 Exeter

Warehouse


14/17

Lecture 7


27

However, there are partial dependencies in STOCK.

ProductNo ProductDesc, Price, PCode,

CDesc.Therefore, these attributes depend on only part ofthe key, i.e. they depend on ProductNo and NOTWHNo ProductNo together.







Stock

28

Second normal form ( 2NF )"all attributes, which are not part of the primary key, must be fullyfunctionally dependent on the whole of the key".

ProductNo ProductDesc Price PCode CDesc

1001 Hammer 10.99 H Hardware

1002 Screwdriver 03.99 H Hardware

1010 Wood Drill 30.49 ET Electrical

4001 Stepladder 49.99 H Hardware

WHNo WHLoc

10 Ilford

30 Exeter

Warehouse

WHNo ProductNo Qty

10 1001 06

10 1002 08

10 1010 10

30 1001 01

30 4001 02

Stock Product


15/17

Lecture 7


29

Second normal form ( 2NF )






WHNo WHLoc

10 Ilford

30 Exeter

Warehouse

WHNo ProductNo Qty

10 1001 06

10 1002 08

10 1010 10

30 1001 01

30 4001 02

Stock Product

Insertion

In PRODUCT,To insert a potentially newproduct code you will need toinsert details of a product,because the key ProductNowould be NULL or empty.This is unacceptable.

30







WHNo WHLoc

10 Ilford

30 Exeter

Warehouse

WHNo ProductNo Qty

10 1001 06

10 1002 08

10 1010 10

30 1001 01

30 4001 02

Stock Product

Deletion

In PRODUCT, if you want tostop selling wood drills,

you need to delete the rowfor wood drill i.e. whereProductNo = 1010.

When you do this, the factthat the product code 'ET'and its description

‘Electrical’ exist, is now lost.


16/17

Lecture 7


31







WHNo WHLoc

10 Ilford

30 Exeter

Warehouse

WHNo ProductNo Qty

10 1001 06

10 1002 08

10 1010 10

30 1001 01

30 4001 02

Stock Product

Editing

In PRODUCT, change thecode description forstepladders from ‘Hardware’to ‘Paint’.

The product code ‘H’ willhave a description of ‘Paint’for this row, and adescription of ‘Hardware’everywhere else. This leadsto inconsistency in the data.

32

ProductNo PCode CDesc

CDesc is not dependent on ProductNo. Its havingtransitive dependecies with ProductNo

Normalisation requires that more new entitytypes be formed to remove the transitivedependencies, but that their key (PCode in thiscase) must also be retained in the original entitytype

Transitive dependencies


17/17

Lecture 7

33

Third normal form ( 3NF )"Entity types are in 3NF when all attributes which are not part of the key are

functionally dependent on the key, the whole key and nothing but thekey".

PCode CDesc

H Hardware

ET Electrical

WHNo WHLoc

10 Ilford

30 Exeter

WarehouseWHNo ProductNo Qty

10 1001 06

10 1002 08

10 1010 10

30 1001 01

30 4001 02

Stock

Product

Product_Code

ProductNo PDesc Price PCode

1001 Hammer 10.99 H

1002 Screwdriver 03.99 H

1010 Wood Drill 30.49 ET

4001 Stepladder 49.99 H

34

Summary NormalisationUnnormalized relation

1NF

2NF

3NF

Remove repeatinggroups

Remove partialdependencies

Remove transitivedependencies

Lecture 8 Normalization

Documents