On Thu, 12 Jul 2018 12:59:15 +0200
Tomas Vondra wrote:
>
>
> On 07/12/2018 03:44 AM, Yugo Nagata wrote:
> > On Wed, 11 Jul 2018 14:37:46 +0200
> > Tomas Vondra wrote:
> >
> >>
> >> On 07/11/2018 01:28 PM, Ashutosh Bapat wrote:
> >
> >>> I don't think we should change add_path() for this. We
On 07/12/2018 03:44 AM, Yugo Nagata wrote:
On Wed, 11 Jul 2018 14:37:46 +0200
Tomas Vondra wrote:
On 07/11/2018 01:28 PM, Ashutosh Bapat wrote:
I don't think we should change add_path() for this. We will
unnecessarily check that condition even for the cases where we do not
create index
On Wed, 11 Jul 2018 14:37:46 +0200
Tomas Vondra wrote:
>
> On 07/11/2018 01:28 PM, Ashutosh Bapat wrote:
> > I don't think we should change add_path() for this. We will
> > unnecessarily check that condition even for the cases where we do not
> > create index paths. I think we should fix the ca
On 07/11/2018 01:28 PM, Ashutosh Bapat wrote:
On Wed, Jul 11, 2018 at 11:03 AM, Yugo Nagata wrote:
Hi,
I found that there is a situation that even when index only scan can be
effective,
the planner doesn't select this. The planner makes indexe paths in descending
order of indexrelid, and t
On Wed, Jul 11, 2018 at 11:03 AM, Yugo Nagata wrote:
> Hi,
>
> I found that there is a situation that even when index only scan can be
> effective,
> the planner doesn't select this. The planner makes indexe paths in descending
> order of indexrelid, and the new path is discarded if its cost is
Hi,
I found that there is a situation that even when index only scan can be
effective,
the planner doesn't select this. The planner makes indexe paths in descending
order of indexrelid, and the new path is discarded if its cost is not less than
the existing paths' cost. As a result, IndexOnlySc
