Normalize the following relation ( emp_id, emp_name, emp_zip, emp_state, emp_city, emp_district ) to 3NF

First Normal Form(1NF)

Here we consider emp_id as the Primary Key and all the columns have atomic values. Thus, the table is in 1NF.

R1

emp_id	emp_name	emp_zip	emp_state	emp_city	emp_district
1001	Hari	282005	state7	Dipayal	Doti
1002	Keshab	555780	state2	Nepalgunj	Banke
1003	Pawan	822005	state7	Dhangadhi	Kailali
1004	Aardi	772290	state4	Pokhara	Kaski
1005	sunita	885009	state3	Kathmandu	Ktm

 

Second Normal Form(2NF)

There aren't any Partial Dependencies in R1 as emp_id is a Single Primary Key. Hence, R1 is also in Second Normal Form(2NF).

 

Third Normal Form(3NF)

Let's check if R1 has Transitive Dependency.

There are multiple transitive dependencies caused by the column emp_zip.

 

Hence we decompose R1 into R2 and R3 to get rid of them.

R2

emp_id	emp_name	emp_zip
1001	Hari	282005
1002	Keshab	555780
1003	Pawan	822005
1004	Aardi	772290
1005	sunita	885009

 

R3

emp_zip	emp_state	emp_city	emp_district
282005	state7	Dipayal	Doti
555780	state2	Nepalgunj	Banke
822005	state7	Dhangadhi	Kailali
772290	state4	Pokhara	Kaski
885009	state3	Kathmandu	Ktm

 

The Transitive Dependency caused by emp_zip is now removed.

However, R3 could be decomposed further into R4 and R5 because of the following transitive dependency:

 

R4

emp_zip	emp_city	emp_district
282005	Dipayal	Doti
555780	Nepalgunj	Banke
822005	Dhangadhi	Kailali
772290	Pokhara	Kaski
885009	Kathmandu	Ktm

 

R5

emp_district	emp_state
Doti	state7
Banke	state2
Kailali	state7
Kaski	state4
Ktm	state3

Here, another transitive dependency (emp_zip → emp_district → emp_state) is removed. Hence, R2, R4 and R5 are in Third Normal Form(3NF).

 

Note:

If we are talking strictly about the theoretical aspect of Normalization, then there is a 'transitive dependency' because if the data is only from Nepal, then the district can derive the state name. In that sense, one should decompose R3 to R4 and R5 to get rid of this transitive dependency.

However, depending upon the nature of data and the kind of application, some database designers may choose to keep it to R3. This is because zip code, district, city, and state belong to the same address unit and are generally fetched together. Thus keeping them in separate tables will unnecessarily cause JOIN operations which are expensive processes in DBMS.