Re: Planner regression in 9.1: min(x) cannot use partial index with NOT NULL

On Mon, Mar 21, 2011 at 10:25 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Robert Haas <robertmhaas(at)gmail(dot)com> writes:
>> On Mon, Mar 21, 2011 at 7:17 AM, Marti Raudsepp <marti(at)juffo(dot)org> wrote:
>>> I know that the Merge Append patch required some changes in the
>>> min/max optimization, which is probably the cause.
>
>> Yeah, I think this is a direct result of commit
>> 034967bdcbb0c7be61d0500955226e1234ec5f04.
>
> Yeah, looks that way.  I'm not sure what it would take to re-support
> this case without losing the other advantages of the change.  Personally
> I'm not terribly excited about it: I don't think that suppressing nulls
> from an index this way is really very useful.  Using a partial index
> probably eats more planner cycles than you'll save, overall.

If only 1% of the table has non-NULL values in that column, maybe not.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company