On Thu, 28 Aug 2003, Tom Lane wrote:
> Stephan Szabo <sszabo(at)megazone(dot)bigpanda(dot)com> writes:
> > ... that would only help if the unique constraint were created before the
> > references constraint.
>
> But in any case, pg_dump ought to be careful to dump index creation
> commands before REFERENCES constraints, I should think. If it isn't
> doing that, there's a performance hit even if the code manages to work.
Good point, the checks would be even slower without the indexes.