Database Normalisation

From TRCCompSci - AQA Computer Science
Revision as of 15:58, 22 December 2016 by ThomasWard (talk | contribs)
Jump to: navigation, search

An un-normalised database is a database that has not undergone the process to remove unnecessary or repeating information from the system's tables. Normalisation will remove this data providing a more efficient database requiring less storage.

To use Normalisation to create a Relational Database we do the following:

Start with the data in one large (un-normalised) table

Remove repeating groups to give 1NF

Remove partial dependencies to give 2NF

Remove non-key dependencies to give 3NF

The final normalized table will have no repeating groups and will have all the tables in the database linked via common keys.