Fourth Normal Form (4NF)
Fourth Normal Form (4NF)
A relation (table) is said to in 4NF:
 If it is in BCNF.
 There must be no multi valued dependency.
Multi valued dependency :
Suppose a relation R (A B C), If multi valued dependency exists between A and B, then it
is denoted by A
B, It means for a single value of attribute A, there are multiple values
of attribute B and there must be B and C independent to each other.
Note: There must be at least three attributes to occur multi valued dependency.
Ex: Consider the following table:
stu_id
stu_mobile
stu_course
101
89832
Java
101
98302
Python
102
45698
PHP
102
45698
Python
103
92342
Java
103
98364
Java
One Student can have multiple mobiles and One student can enroll on multiple courses. The
Candidate key in above table {stu_id, stu_mobile, stu_course} The table stratifies the rule of BCNF
[The whole attribute in the table are the part of the candidate key] But the table still have data
redundancy due to multi valued dependencies:
stu_id
stu_mobile [ for stu_id 101 and 103, there are more than one mobile]
stu_id
stu_course [ for stu_id 101 and 102, there are more than one course ]
stu_mobile and stu_course are also independent to each other.
Therefore, to satisfy 4NF, it needs to be decompose into
R1
(stu_id,
stu_mobile)
[The multi valued dependent columns
are moved into separate table.]
R2(stu_id, stu_course)
See how data redundancy is removed by decomposing it into 4NF:
R1(stu_id, stu_mobile)
R (stu_id, stu_mobile, stu_course)
stu_id
stu_mobile
stu_id
stu_mobile
stu_course
101
89832
101
89832
Java
101
98302
101
98302
Python
102
45698
102
45698
PHP
103
92342
102
45698
Python
103
98364
103
92342
Java
103
98364
Java
R2 (stu_id, stu_course)
stu_id
101
101
102
102
103
stu_course
Java
Python
PHP
Python
Java
 Multi valued dependency exist:
o {stu_id
stu_mobile, stu_id
stu_course
o stu_mobile and stu_course are indepdendent
 Data Redundancy:
o For every new course, mobile needs to be repeated.
o For every new mobile, course needs to be repeated.
5NF:
A relation in the DBMS is in the Fifth Normal Form if it is in the Fourth Normal Form and does not consist of any join
dependency. The joining must also be lossless for this.
Fifth Normal Form is satisfied when all the tables present in the DBMS are broken down into as many more tables as possible to
ensure that there is no redundancy.
Fifth Normal Form is also commonly known as PJ/NF or Project Join Normal Form.
Example
SUBJECT
LECTURER
SEMESTER
DBMS
Manish
1
DBMS
Anil
1
JAVA
GP
1
JAVA
NK
2
HTML
AK
1
In the table above, John takes lectures for both, math as well as computer in the first semester but does not take math in the
second. In such a case, a combination of all fields in the table is needed to identify the data validity.Now assume that we have to
add another semester, where we do not know the subject or the lecturer. What do we do then? This is where Fifth Normal Form
comes in. We divide the table into smaller tables
Table1
Semester
Subject
1
DBMS
1
JAVA
1
HTML
2
JAVA
Table2
Subject
Lecturer
DBMS
Manish
DBMS
Anil
JAVA
GP
JAVA
NK
HTML
AK
Table3
Semester
Lecturer
1
Manish
1
Anil
1
GP
2
NK
1
AK