Re: identity columns

On Wed, 2 Jan 2002 22:46:54 -0500, "John" <ru_re_re_move_zel_at_ecl_re_re_move_ipse.net> wrote:

>There seems to be a lot of talk in this NG about the good, bad and ugly
>aspects of primary keys - especially Identity columns.
>
My apologies for snipping so much of your message, but it is for brevity only - no offence meant.

I am a little like Pooh Bear on this topic (I hold different opinions on different days of the week). Theoretically, key fields should not have data content - so sayeth Ted Codd - so I often use a simple sequential number. On the other hand, if you happen to have a unique data field (which, IMHO, an email address is not) then I see no reason not to use it as a key. On other days, I will even concatenate two non-unique fields to make one unique one especially so that it can be made into a key (especially when I am working on Pick-like databases).

IMHO, this is one area where practicality rules, not theory.

-- 
Mark A. Preston, The Magpie's Nest, Lancashire, UK
Email   : mark_at_magpiesnest.co.uk
Website : www.magpiesnest.co.uk