Re: Surrogate Key vs Production Key

DA Morgan <damorgan_at_x.washington.edu> wrote in message news:<1099010940.913494_at_yasure>...

> I hopefully never did. I said that to maintain the integrity of the
> data one must then create additional constraints. Sorry if my original
> posts were unclear. I often type these missives after a full 10 hour
> work day followed by 3 hours of lecture.

:) Same here. Up at 5:30, followed by 9 hours solid work and 2 hours getting aggravated by traffic, then 2 hours reading building laws to find out how the builder is stuffing up my house, then hopefully some time here for a quick read and a chat, then off again reviewing book proposals for another coupla hours, then the dog wants to go out and the cat tricks me to get out as well, then 10 minutes chasing both of them and finally just enough time to undress before shuteye kicks in. Endlessly. I need a boat. Even a canoe will do...

> I am refering to your unique index to maintain data integrity. It will
> allow NULLs if you don't also add check constraints.

Of course. And the overhead of a NOT NULL check is absolutely minimal.

> John Smith, even by the wildest of imaginations is not a natural key.
> Try your example again with a real one such as an ISBN or Postal Code.

It's not my example. Someone else brought him in. I'd NEVER use a name as a PK, natural or not.

>
> Look more closely. ;-)

Need better reading glasses, I think... :) Received on Fri Oct 29 2004 - 05:48:22 CDT