On 16 March 2016 at 12:58, Robert Haas <robertmhaas@gmail.com> wrote:
> On Tue, Mar 15, 2016 at 6:55 PM, David Rowley
> <david.rowley@2ndquadrant.com> wrote:
>> On 16 March 2016 at 11:00, Robert Haas <robertmhaas@gmail.com> wrote:
>>> I don't see why we would need to leave aggpartial out of the equals()
>>> check. I must be missing something.
>>
>> See fix_combine_agg_expr_mutator()
>>
>> This piece of code:
>>
>> /*
>> * Aggrefs for partial aggregates are wrapped up in a PartialAggref,
>> * we need to look into the PartialAggref to find the Aggref within.
>> */
>> foreach(lc, context->subplan_itlist->tlist)
>> {
>> PartialAggref *paggref;
>> tle = (TargetEntry *) lfirst(lc);
>> paggref = (PartialAggref *) tle->expr;
>>
>> if (IsA(paggref, PartialAggref) &&
>> equal(paggref->aggref, aggref))
>> break;
>> }
>>
>> if equals() compared the aggpartial then this code would fail to find
>> the Aggref in the subnode due to the aggpartial field being true on
>> one and false on the other Aggref.
>
> ...and why would one be true and the other false?
One would be the combine aggregate (having aggpartial = false), and
the one in the subnode would be the partial aggregate (having
aggpartial = true)
Notice in create_grouping_paths() I build a partial aggregate version
of the PathTarget named partial_group_target, this one goes into the
partial agg node, and Gather node. In this case the aggpartial will be
set differently for the Aggrefs in each of the two PathTargets, so it
would not be possible in setrefs.c to find the correct target list
entry in the subnode by using equal(). It'll just end up triggering
the elog(ERROR, "Aggref not found in subplan target list"); error.
-- David Rowley http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training & Services
