The database system has three models, i.e., conceptual database, logical database
and physical database. Conceptual database models are for each view of real world
problems, e.g., Entity Relationship (ER) model. Logical database deals with the
construction of data model, e.g., relational database schema. Physical database describes
file organization and indexes design. A clear distinction is made between the logical
and physical model of database. It is the logical view that is clearly understood by the
users who are not database developers. Data is stored in tables which constitutes
attributes. Each attribute belongs to a particular domain. Attributes that are in the
same column are said to be members of a set. Attributes in the same row are members
of an ordered n-tuple. A relation is formed by n-tuples in the table. One or more
columns form a key of the table that uniquely determines each relation (Rocket U2,
2010). Relational database design, like database design using any other data model, is
far from being a completely automated process in the current state of database
technology. It is an activity that requires the close attention of the database designer,
who may be one individual, example DBA, or a team working with DBA. This activity
consists of identifying that portion of the enterprise for which the database application
is being designed: the entity sets, their attributes, the domain on which attributes are
defined, and the constraint that these attributes have to satisfy. Then the design of
the relational scheme can begin (Ramez and Shamkant, 2007). Two approaches are
generally used in designing a relational database, the decomposition approach and
the synthesis approach. The decomposition approach starts with universal relation
and the associated set of constraints in the form of functional dependencies,
multivalued dependencies and join dependencies, whereas the synthesis approach
starts with a set of functional dependencies on a set of attributes.
|