Re: atomic

From: Bob Badour <bbadour_at_pei.sympatico.ca>
Date: Fri, 02 Nov 2007 18:23:05 -0300
Message-ID: <472b953e$0$14867$9a566e8b@news.aliant.net>

paul c wrote:

> Bob Badour wrote:
>

>> paul c wrote:

>
> ...
>

>>> I still wonder why one would want both tuples to apply to the same 
>>> pizza.
>>
>> With your predicate, I doubt one would, which is why one would declare 
>> a candidate key that would reject the relation.

>
> I agree with that and with Roy H who I think has said the same thing.
> I ask myself how many ways could such a relation arise and the
> answer seems to be only one: because somebody defined it and populated
> it, eg., at least I don't think it could be produced with projection and
> group operators from some other relation(s). Not saying there isn't
> another way, just that I don't see one.

If one has an operation that groups relations, then it's easy to declare an expression that evaluates as an RVA--even an empty RVA.

> So for the examples so far, it seems that even if they aren't
> contradictory according to theory, I'd say they are at least a bit wacky
> as far as their predicates are concerned, unless a primary key is
> specified that doesn't include the "set" or RVA attributes.

Let's turn the predicate around:

Price Toppings
----- ========

$10      { cheese, pepperoni, mushroom }
$5       {}

I used double-underscore to indicate the primary key. Received on Fri Nov 02 2007 - 16:23:05 CDT