Pragmatically speaking, rather than declaring NULLs to be anathema, it makes more sense to simply avoid them as much as possible. If it looks like you're going to have a lot of NULLs, you probably need a separate table for the field. But if unknown values are a rare exception, it may be both easier and neater to use NULLs. Why did Codd make systematic handling of NULLs a mandatory requirement for an RDBMS, if he didn't intend that they should ever be used for anything?

In practice, it probably makes more sense to think of normalization as a continuum rather than a binary state. This database is MORE normalized, that one is LESS. Purists probably disagree, but I tend to think in terms of what's going to be used in the real world. If you're able to convince a user to split their flat file up, consider that a victory, and don't bemoan the fact that it's not in 3NF.

SOLUTION

Jankovsky

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial