Re: constraints in algebra instead of calculus
From: David Portas <REMOVE_BEFORE_REPLYING_dportas_at_acm.org>
Date: Mon, 21 May 2007 07:04:53 +0100
Message-ID: <YoqdnTkzb60Xq8zbnZ2dnUVZ8q6unZ2d_at_giganews.com>
> I've never seen anybody justify why a foreign key must be a candidate key
> in the referenced relation/relvar/table except for 'so-and-so says so'.
> Whereas I've seen examples where the non-key reference is essential.
Date: Mon, 21 May 2007 07:04:53 +0100
Message-ID: <YoqdnTkzb60Xq8zbnZ2dnUVZ8q6unZ2d_at_giganews.com>
"paul c" <toledobythesea_at_oohay.ac> wrote in message news:JQ94i.200044$DE1.108062_at_pd7urf2no...
>
> I've never seen anybody justify why a foreign key must be a candidate key
> in the referenced relation/relvar/table except for 'so-and-so says so'.
> Whereas I've seen examples where the non-key reference is essential.
I think the main answer is "history". As I see it there is no real restriction. The term "foreign key" is merely a shorthand or syntactical device meaning the type of constraint you described. That does not preclude other types of constraints - it's just that by convention they are not "foreign keys".
-- David PortasReceived on Mon May 21 2007 - 08:04:53 CEST