On Wed, 19 Jan 2005, Stephan Szabo wrote:
>
> On Wed, 19 Jan 2005, Dann Corbit wrote:
>
> > True, but the standard says nothing about the creation of an index, so
> > you can make it behave in any way that you see fit.
>
> The unique index is however used to model the unique constraint in
> PostgreSQL which I had thought was clear from my statement so giving the
> unique index behavior which makes it unable to model the constraint
> wouldn't be terribly workable (without rewriting the constraint to be
> modeled in a separate fashion).
Actually, thinking about it, I think I generally disagree with the idea
that unique is the correct word to use in any case. It's used in both the
constraint and in the predicate to have its current meaning (as described
in my earlier message) and I think the word distinct already the semantics
of treating two NULLs in the fashion such that an index wouldn't allow
two.
