Re: Relation Schemata vs. Relation Variables

From: Bob Badour <bbadour_at_pei.sympatico.ca>
Date: Fri, 25 Aug 2006 02:21:08 GMT
Message-ID: <o8tHg.1220$9u.16561@ursa-nb00s0.nbnet.nb.ca>

Marshall wrote:

> Bob Badour wrote:
>

>>The concept of a candidate key is pretty much meaningless
>>with respect to relations.

>
> Mmm, I'm going to have to disagree with that one. It's meaningless
> as a prescriptive constraint, yes. But it still has descriptive value.

The most specific type is interesting. I guess the concept of a candidate key for the MST has some meaning.

>>What is irreducible for a relvar may very well be reducible
>>for the relvar's current value.

>
> This one I agree with.

I prefer to think in terms of types. The MST of the relation must be a subtype of the declared type of the relvar. A candidate key is part of the type declaration for a relvar. Received on Thu Aug 24 2006 - 21:21:08 CDT