Tuesday, 23 July 2013

SQL SERVER - NORMALIZATION With Examples















What is Normalization ?
  Normalization is a process of eliminating Redundant data and storing the related information in a table.


1. Eliminating Redundant data.
2. Faster update
3. Improve performance
4. Performance in indexes

Let we see different Normalization forms

1. First Normal Form (1NF)
    If a Table is said to be 1NF then it should satisfy following rules.

  • Each cell must have one value
  • Eliminating Duplicate Columns
  • Create a separate table for group of related data and each row must be identify by primary key.
That means each cell must have single value and each row should be uniquely identified by Primary key

For Example :

Name
Department
Phone Number
Rajesh
Computer
3452342,1234563,2345612
Suresh
Electronics
2398521,2323177,5302994
Praba
Civil
3958218
In the above we can see the duplicate columns phone numbers have more than one value , we have to eliminate that and create a group of related data with Unique row identification by specifying a primary key for the table

Rule 1. By applying above rule each cell must have one value above table changes like below

Name
Department
Phone Number
Phone Number
Phone Number
Rajesh
Computer
3452342
1234563
2345612
Suresh
Electronics
2398521
2323177
5302994
Praba
Civil
3958218



Rule 2 & 3 . By applying second rule and third rule no more duplicate columns and each row must be unique is applied     to above table.

Id
Name
Department
Phone Number
1
Rajesh
Computer
3452342
2
Rajesh
Computer
1234563
3
Rajesh
Computer
2345612
4
Suresh
Electronics
2398521
5
Suresh
Electronics
2323177
6
Suresh
Electronics
5302994
7
Praba
Civil
3958218


2. Second Normal Form (2NF)
    The Table must be in second normal form , Then it should satisfy the following rules.
  •  It should satisfy first normal form
  •  Separate the particular columns ,values are duplicated in each row  should be place in separate table
  •  Create the relationship between the tables
From the above table we can see the column name and department are repeated in each row ,This two columns can be maintained in another table and make a relationship between these two tables 

EmpId
Name
Department
1
Rajesh
Computer
2
Suresh
Electronics
3
Praba
Civil

Id
EmpId
PhoneNumber
1
1
3452342
2
1
1234563
3
1
2345612
4
2
2398521
5
2
2323177
6
2
5302994
7
3
3958218
In the above table Empid is played as Primary key for the first table and foreign key for the second table.



3. Third Normal Form (3NF)
     The table must be in 3NF,if it is satisfying the following rules
  •  Must be in 2NF
  •  Separate the columns that are not dependent upon the primary key of the table.
Product
Price
Tax
LED
23000
20%
AC
15000
10%
Fridge
12000
15%

From the above table you can see that Tax Column is not dependent on Product Primary key column, It is dependent on Price so we separate that in to two different table.

Product
Price
LED
23000
AC
15000
Fridge
12000

Price
Tax
23000
20%
15000
10%
12000
15%


4. Fourth Normal Form (4NF)
  • It should be in 3NF
  • The non key columns should be dependent on full primary key instead of partial key , If then separate it.
From the following table "EmployeeName" Non-Key column not dependent on full primary key "ManagerId,EmployeeId,TaskID" it depends upon the EmployeeId  Partial Key so it can be separated.


ManagerId
EmployeeId
TaskID
EmployeeName
M1
E1
T1
Rajesh
M2
E1
T1
Rajesh

ManagerId
EmployeeId
TaskID
M1
E1
T1
M2
E1
T1

EmployeeId
EmployeeName
E1
Rajesh

That's it from this article we can see the normalization and there concepts Fully.