Bruno Wolff III [mailto:bruno@wolff.to] Wrote
> I think the real problem is that check constraints on tables
> aren't used by the optimizer. Given that, what you have below
> is expected.
> There has been talk about that in the past, but I haven't
> heard anything recently about someone considering implenting that.
>
> For your problem consider not using a partial index. It isn't
> going to save anything if it has a constraint matching that
> of the table.
Ahh, I get it now,...
If a column has a CHECK (col IN (1,2,3)) and a query says .. WHERE col
= 4; then the planner should know that the query will return 0 rows, right?
... John