Normalize The Relation:
Once a relation has been defined for an entity,
it should be analyzed according to the normalization rules. For example, see the
'STUDENT' relation as shown below.
STUDENT (Roll-No, Name, Marks, Address,
City, ZipCode)
Is it in Domain-Key Normal Form, (DK/NF)? For
this purpose, you have to know the constraints on this relation. Most of the
constraints are discovered just from the names of the attributes and knowledge
about the nature of the business such as;
- First, Roll-No determines all of the other
attributes, because values of other attributes of a specific row can be
determined from a given value of Roll-No.
- There are other constraints that occur due
to functional dependencies. ZipCode determines city and province etc. For
this purpose, we have to divide the 'STUDENT' relation into two relations
such as STUDENT (Roll-No, Name, Marks) relation and ADDRESS relation. In this
case, there will be no modification anomalies. In short, we can say that to
represent an entity into relation, we have to perform the following steps.
1. Transform an entity into relation that has all of the entity's attributes
as columns.
2. Normalize the relation. In many cases, the design can be improved by
developing a set of relations in
Example:
Transformation of university E-R data model into
relational data model.

|