Re: Normalization, Natural Keys, Surrogate Keys
Date: Thu, 23 May 2002 18:12:20 GMT
Bernard Peek wrote:
> >In reflection I tend to agree. By using an artificial surrogate key -
> >aside from making the database easier to read - we avoid the problem
> >of potenially changing data.
> I think that's incorrect. By using an artificial surrogate key you avoid
> detecting changed data, which is subtly different. If there is a genuine
> natural key then it is impossible for it to change. The key uniquely
> identifies one and only one instance of an entity. If the key changes it
> can only be because the instance it refers to has changed, and in that
> situation you want to detect the change.
By this standard it seems that the only natural key for a human being is their entire DNA sequence. Oh wait, clones are just around the corner, nevermind. Seriously though, I'm not following your meaning here. Can you provide an example of something and it's never changing natural key?
Richard Received on Thu May 23 2002 - 20:12:20 CEST