Relations create the database. They must satisfy various properties, like no duplicate tuples, no ordering of tuples, and atomic attributes, etc. Relations that satisfy these basic needs may still have some undesirable characteristics such as data redundancy, and anomalies.
What are these undesirable characteristics and how can we eliminate them from the database system? This part is an attempt to answer this question. Though, please note that most of these undesirable properties do not arise if the database modelling has been carried out very carefully using some method like the Entity-Relationship Model. It is still important to use the methods in this unit to check the database that has been acquired and ensure that no mistakes have been made in modeling.
The central thought in these discussions is the concept of Database integrity, the notion of functional dependency, which rely on understanding the semantics of the data and which deals with what information in a relation is dependent on what other information in the relation. Our prime objective in this section is to describe the concept of data integrity, functional dependence and then describes normal forms using functional dependencies and using other types of data dependencies.
OBJECTIVES
After reading this section you should be able to
- Explain the concept of entity integrity;
- Define relational database referential integrity constraints;
- Explain the concept of functional dependency, and
- show how to use the dependencies information to decompose relations; whenever essential to obtain relations that have the desirable properties that we want without losing any of the information in the original relations.