# Re: Constraints and Functional Dependencies

Date: Sat, 24 Feb 2007 20:47:55 GMT

Message-ID: <%v1Eh.2422$kr6.1867_at_trndny09>

"Marshall" <marshall.spight_at_gmail.com> wrote in message
news:1172304086.527354.319940_at_j27g2000cwj.googlegroups.com...

> I hope no one will mind too much if I take a break from our

*> usual sort of discussion and talk about database theory for
**> a bit. I feel like free associating about functional dependencies.
**>
**> I like to consider a system under which there is only one type
**> of constraint: a formula of first-order logic. That is to say,
**> there are no uniqueness constraints, foreign keys, primary keys,
**> etc., except as such can be expressed as FOL formulas.
**>
**> In a data management context, there is some value to restricting
**> what we can quantify over as being only attributes of declared
**> relations (whether variables or constants.) So, we can't express
**> the "no-upper-bound" property of the natural numbers; they aren't
**> a database table so we can't quantify over them.
**>
**> With such a system, a relation R with attribute a (which I will
**> write as R(a)) having a as a foreign key into S(b) is expressed
**> as follows:
**>
**> forall R(a): exists S(b): a = b
*

Nitpicking response: the above not only expresses the foreign key concept,
but also referential integrity.

I don't want to make a mountain out of a mole hill here. I just want to
reach consensus, if possible, before considering the rest of your
discussion.
Received on Sat Feb 24 2007 - 21:47:55 CET