Re: WIP: Upper planner pathification

On 3 March 2016 at 04:29, Alvaro Herrera <alvherre(at)2ndquadrant(dot)com> wrote:
> Simon Riggs wrote:
>
>> ISTM that we are clearly "going for it"; everybody agrees we should apply
>> the patch now.
>>
>> The longer we hold off on applying it, the longer we wait for dependent
>> changes.
>
> Agreed -- we need this in tree as soon as realistically possible.
>
> There is a a bit a problem here, because this patch conflicts heavily
> with at least one other patch that's been in the queue for a long time,
> which is Kommi/Rowley's patch for parallel aggregation; the more we
> delay applying this one, the worse the deadlines for that one.
>
> I assume they are hard at work updating that patch to apply on top of
> Tom's patch. It's not realistic to expect that we would apply any
> further planner changes before this one is in.

I agree that it would be good to get this in as soon as possible. I'm
currently very close to being done with writing Parallel Aggregate on
top of the upper planner changes. So far this version is much cleaner
as there's less cruft added compared with the other version, of which
would need to be removed again after the upper planner changes are in
anyway. Putting parallel aggregate in first would be asking Tom to
re-invent parallel aggregate when he rebases the upper planner stuff
on the new master branch, which makes very little sense.

So I agree that it would be nice to get the upper planner changes in
first, but soon! not at the end of March 'fest, as doing so would most
likely kill parallel aggregate for 9.6, and I kinda think that would
be silly as (I think) it's pretty much the biggest missing piece of
the parallel query set.

--
David Rowley http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services