Re: Normalization, Natural Keys, Surrogate Keys
Date: Mon, 20 May 2002 18:04:05 GMT
In article <3CE583F3.7F5D20D5_at_exesolutions.com>,
> If I didn't get paid so darned much money dumping surrogate keys and cleaning
> up the messes I'd actually be upset. And not once has there been a resulting
> performance problem that warranted recreating a surrogate key and instituting
> a fake primary key using a unique index with a not null constraint.
The SSN is a surrogate key itself, and a BAD one at that. There are duplicates out there. It is "just as good" as a surrogate. In fact, it might be worse, because at least I can QA my surrogate, but I can't QA the government's surrogate for Person.
Secondly, as someone else explained, the error you illustrate of duplication is NOT due to the use of a surrogate key, but of bad usage of integrity operations (checks, triggers or stored procedures) that would have prevented it.
(Any opinions expressed are strictly mine only and not my employer's)
Paul Tiseo, Intermediate Systems Programmer Reply To: 123tiseo.paul_at_123mayo.edu (drop the numbers) Received on Mon May 20 2002 - 20:04:05 CEST