On 6/17/05, Marc G. Fournier <scrappy@postgresql.org> wrote:
> On Fri, 17 Jun 2005, Tom Lane wrote:
>
> > "Marc G. Fournier" <scrappy@postgresql.org> writes:
> >> Does that make sense? Would it ever get used?
> >
> > It could get used if one of the two values is far less frequent than the
> > other. Personally I'd think about a partial index instead ...
>
> Hrmmmm, hadn't thought of that ... wouldn't you have to build two indexes
> (one for true, one for false) for this to be completely effective? unless
> you know all your queries are going to search for one, but not the other?
>
I guess it will be effective only if you know wich value will be less
frequent... on the other value a sequential scan will be a win, isn't
it?
--
regards,
Jaime Casanova
(DBA: DataBase Aniquilator ;)
