Re: No Index-Only Scan on Partial Index

On Oct 2, 2013, at 5:07 AM, Merlin Moncure <mmoncure(at)gmail(dot)com> wrote:

> > Hrm. I get a seq scan for that query:
> >
> > create index on try(upper_inf(irange));
> > explain select * from try where upper_inf(irange);
> > QUERY PLAN
> > -----------------------------------------------------------
> > Seq Scan on try (cost=0.00..1887.00 rows=33333 width=68)
> > Filter: upper_inf(irange)
> >
> > True also if I just select the irange. Is the filter the issue, here?
>
> Turn off seq scan...

That rewards me with a bitmap heap scan:

EXPLAIN select * from try where upper_inf(irange);

QUERY PLAN
--------------------------------------------------------------------------------------
Bitmap Heap Scan on try (cost=935.63..2197.63 rows=33333 width=68)
Filter: upper_inf(irange)
-> Bitmap Index Scan on try_upper_inf_idx (cost=0.00..927.30 rows=50000 width=0)
Index Cond: (upper_inf(irange) = true)

But anyway, I still don’t understand why, if the function used to store the value is immutable (as upper_inf() is), why Postgres doesn't do an index scan. Is this something that could be improved in the planner?

Thanks,

David