On Monday 02 Jun 2003 3:00 pm, erwan ancel wrote:
> well, no... these are not direct foreign keys. The constraint here is
> that for a given record of D, B pointed by A pointed by the given D must
> be the same as B pointed by C pointed by the given D.
>
> This is not a foreign key, or foreign keys are much more than what I
> thought.
> Erwan
Sounds like it might be the scenario described in:
http://techdocs.postgresql.org/guides/BriefTriggerExample
--
Richard Huxton