Joseph Meehan said:In short it means dividing date into different related tables in a
logical way.
For example a retailer may list the name address and phone number of
each customer as well as a list of each item purchased.
You could put all that into a single table repeating the name address
and phone number for each purchase. This is going to cause the database to
grow very large and slow in no time. The proper way is to normalize the
table into two tables. A customer table will have each customer listed
along with their demographics like address. This table will be linked to a
second table listing all sales. The link between tables will allow listing
all the sales made to a particular customer or all the customers who bought
a specific item etc.
That was a simple case, in real life sometimes the design of normalized
tables can be as much of an art form as a science. In any case it is
critical for proper performance of all but the most simple databases.
Arvin Meyer said:Normalization in a relational database is as much a study in knowing whe not
to normalize as when to do it. A good priner of when to normalize is
available here:
http://www.microsoft-accesssolutions.co.uk/database_normalization_basics.htm
Relational rules are based on Relational Set Theory in Calculus and have a
specific basis. The experience in interpreting those rules and applying them
to business situations are what sets database developers apart from other
developers and programmers. Knowing *when* to break those rules is also a
mark of experience, too often confused by novice developers.
--
Arvin Meyer, MCP, MVP
Microsoft Access
Free Access Downloads
http://www.datastrat.com
http://www.mvps.org/access