Automatic Normalisation
Date: Wed, 22 Oct 2008 17:07:39 -0700 (PDT)
Message-ID: <47fa5b61-b170-4509-a5aa-688bdfbd3b8f_at_c36g2000prc.googlegroups.com>
The rules governing whether a given relvar coupled with dependency rules are in a normal form are well defined.
Unfortunately, many of the FDs are actually business rules in the designer's head, not easily expressed in the form FD A -> B etc. The easiest way to normalise data is (still) using pen and paper AFAIK! Possibly sketching an ERD or something.
If you can list the attributes you need to track, and if you can identify the attributes' inter-dependencies, it should be possible to deduce any desired NF mechanically from the rules.
I have searched the forum (and the FAQ, which appears to have been a non-starter) without success. I Googled 'automatic normalisation' and it looks like some people (mainly academics) have had a go at this. My question is, does anybody in this forum know of, of have any experience with software that guides the designer to the correct solution. Are any of them worth trying out?
Alternatively, if there isn't anything, why can't it be done?
Thanks
Received on Thu Oct 23 2008 - 02:07:39 CEST