Re: Normalization by Composing, not just Decomposing
Date: Thu, 8 Apr 2004 15:37:04 -0400
Just because the remedy for a given lack of normalization is generally decomposition doesn't necessarily mean that decomposition is what normal forms are ABOUT.
For each normal form, it's worth asking, what problems come about if one does not adhere to this normal form? And in fact the best treatments of normal forms do precisely that.
For 1NF the problem is an access problem: you have to look in more than one column to find the answer to a single question.
For all the other normal forms, the problem is an update problem. You have to change more than one row to change a single fact, and, if you don't, you could end up with a self contradictory database.
Your previous answers suggest you already know this, so I'm wondering where the rest of your comment is leading.
BTW, the fact that 1NF solves a different problem than all the others could be related to your claim that PICK data is normalized with regard to the 2NF and 3NF rules, and maybe more., but not 1NF.
The fact that normal forms avoid (rather than solve) certain problems also gives a handle on WHEN to normalize: if it ain't broke, don't fix it. Received on Thu Apr 08 2004 - 21:37:04 CEST