Re: Database design/normal form conundrum

From: Laconic2 <laconic2_at_comcast.net>
Date: Sat, 27 Mar 2004 15:18:23 -0500
Message-ID: <0MCdndOA3sTwfPjd4p2dnA_at_comcast.com>


Just because it's the obvious answer, doesn't mean it's the right answer. Simplicity is sometimes in the mind of the beholder.

To me, it is far simpler to scan one column, when looking for data to include in a set, than it is to scan five columns.

The cost in mental complexity of joining two tables is so low as to not be worth discussing.

The cost in computer resources to carry out the join is also not worth discussing until you get up into the terabytes.

The benefits of normalized data far outweigh the costs, in this case, IMO.

I also note that it's fairly straighforward to crosstabulate the data, if you ever want to do that. You can use pivot tables, crosstab queries, or a technique called "turning a table on its side".

"Bernard Peek" <bap_at_shrdlu.com> wrote in message news:Rtphzv4YCDZAFwD+_at_shrdlu.com...
> Remember the KISS principle (Keep It Simple, Stupid.)
>
> If you know for sure that there will always be exactly 15 fields
> required then a single table is the obvious answer.
Received on Sat Mar 27 2004 - 21:18:23 CET

Original text of this message