Florian Pflug <fgp(at)phlo(dot)org> writes:> Anyway, I was wondering why we need guaranteed uniqueness for FK> relationships anyway.
It's required by spec, and the semantics aren't terribly sensiblewithout it.
regards, tom lane