Database Design

Hello-

I recently came across an application that has 400+ tables in an Oracle database. The tables do not have referential integrity contraints between them. For example a record in a child table can exist without a corresponding row in the parent table. Or a data column can contain a value that is not present in the master table containing all different values.

What are the typical reasons for such a database design without referential constraints ?

My guess is that by doing away with the referential contraints, the application performs faster. Further reasoning is that using APIs to create/delete/update data would keep the integrity of the data without a need for the database referential constraints.

I cannot seem to agree with the approach or its reasoning. Without the database referential contraints, the quality of data goes down the drain over a period of time - mostly due to data fixes needed outside of APIs via scripts, etc.

I would like to know your perspective on this aspect..pros/cons and any material that discusses this particular aspect of the database design is much appreciated.

Thanks,
Nan. Received on Fri Jan 02 2004 - 13:26:01 CST