It defines that no primary key value can be null. This is as the primary key is used to recognize individual tuple in the relation. So we will not be able to recognize the records uniquely containing null values for the primary key attributes. This constraint is specified on single individual relation.
Example: Let R be the Table
|
A#
|
B
|
C
|
|
Null
|
B1
|
C1
|
|
A2
|
B2
|
C2
|
|
Null
|
B3
|
C3
|
|
A4
|
B4
|
C3
|
|
A5
|
B1
|
C5
|
Note:
1) '#' shows the Primary key of a relation.
In the relation R above, the primary key has null values in the tuples t1 & t3. NULL value in primary key is not allowed, therefore, relation instance is an invalid instance.