Re: Surrogate primary key plus unique constraint vs. natural primary key: data integrity?

From: Jan Hidders <hidders_at_gmail.com>
Date: Wed, 20 Mar 2013 02:32:21 +0100
Message-ID: <514911a5$0$26880$e4fe514c_at_dreader37.news.xs4all.nl>


On 2013-03-19 15:16:54 +0000, Cimode said:

> > And that's probably why my professor had told us to use natural

>> (composite) keys by default; because then you *have to* think about
>> *proper* unification *and* such integrity issues as the one with the
>> overlapping foreign keys are much less likely to remain unnoticed.

> Unique identification is EOI specific which makes defining *default*
> type of identifier at best a moot concept, at worst a subjective
> choice.

That's why we have conceptual data models (as opposed to logical data models) where you don't have to make that choice. Why does this worry you so much?

  • Jan Hidders
Received on Wed Mar 20 2013 - 02:32:21 CET

Original text of this message