Re: Artificial Primary keys

"Bernard Peek" <bap_at_shrdlu.com> skrev i en meddelelse news:1RWc$oFw4BU8Ew$$_at_shrdlu.com...
> In message <3c4f3e5e$0$13976$edfadb0f_at_dspool01.news.tele.dk>, Jan Emil
> Larsen <jel_at_g-it.dk> writes
>
> >A key should be imutable, and should therefore be without information in
it
> >self.
>
> The first is true but the second doesn't follow from it.

That is right. I goes the other way round: If it has information in it self, it may change.
No-information in the key is a measure to secure immutability.

>You should
> definitely try as hard as you can to have an immutable primary key but
> where possible it should be immutable because the data in it really does
> identify one and only one thing.

Agreed. A key should identify one and only one thing (an entity; or the table at BCNF)

>It's only when you don't have such a
> key that you need an artificial one.

Yes, but that happens quite often. Received on Thu Jan 24 2002 - 19:37:46 CET