> On 25 Mar 2020, at 15:35, Pengzhou Tang wrote:
> Thanks a lot, the patch has a memory leak in the lookup_hash_entries, it uses
> a list_concat there
> and causes a 64-byte leak for every tuple, has fixed that.
>
> Also, resolved conflicts and rebased the code.
While there hasn't been a revie
On Fri, Mar 20, 2020 at 07:57:02PM +0800, Pengzhou Tang wrote:
Hi Tomas,
I rebased the code and resolved the comments you attached, some unresolved
comments are explained in 0002-fixes.patch, please take a look.
I also make the hash spill working for parallel grouping sets, the plan
looks like
and I also need
to
make hash spill work in the final stage of parallel grouping sets, will do
that
tomorrow.
the conflicts mainly located in the handling of hash spill for grouping
sets,
the 0004-reorganise- patch also make the refilling the hash table stage
easier and
can avoid the
To summarize the current state of parallel grouping sets, we now have
two available implementations for it.
1) Each worker performs an aggregation step, producing a partial result
for each group of which that process is aware. Then the partial results
are gathered to the leader, which then
Thanks to reviewing those patches.
Ha, I believe you meant to say a "normal aggregate", because what's
> performed above gather is no longer "grouping sets", right?
>
> The group key idea is clever in that it helps "discriminate" tuples by
> their grouping set id. I haven't completely thought this
On Mon, Feb 3, 2020 at 12:07 AM Richard Guo wrote:
>
> Hi Jesse,
>
> Thanks for reviewing these two patches.
I enjoyed it!
>
> On Sat, Jan 25, 2020 at 6:52 AM Jesse Zhang wrote:
>>
>>
>> I glanced over both patches. Just the opposite, I have a hunch that v3
>> is always better than v5. Here's my
Hi Amit,
Thanks for reviewing these two patches.
On Sat, Jan 25, 2020 at 6:31 PM Amit Kapila wrote:
>
> This is what I also understood after reading this thread. So, my
> question is why not just review v3 and commit something on those lines
> even though it would take a bit more time. It is
Hi Jesse,
Thanks for reviewing these two patches.
On Sat, Jan 25, 2020 at 6:52 AM Jesse Zhang wrote:
>
> I glanced over both patches. Just the opposite, I have a hunch that v3
> is always better than v5. Here's my 6-minute understanding of both.
>
> v5 (the one with a simple partial aggregate)
emented according to different methods, which causes confusion.
> > >
> >
> > Both the idea seems to be different. Is the second approach [1]
> > inferior for any case as compared to the first approach? Can we keep
> > both approaches for parallel grouping sets, if
the idea seems to be different. Is the second approach [1]
> inferior for any case as compared to the first approach? Can we keep
> both approaches for parallel grouping sets, if so how? If not, then
> won't the code by the first approach be useless once we commit second
>
pared to the first approach? Can we keep
both approaches for parallel grouping sets, if so how? If not, then
won't the code by the first approach be useless once we commit second
approach?
[1] -
https://www.postgresql.org/message-id/CAN_9JTwtTTnxhbr5AHuqVcriz3HxvPpx1JWE--DCSdJYuHrLtA
I realized that there are two patches in this thread that are
implemented according to different methods, which causes confusion. So I
decide to update this thread with only one patch, i.e. the patch for
'Implementation 1' as described in the first email and then move the
other patch to a separate
On Sun, Dec 1, 2019 at 10:03 AM Michael Paquier wrote:
> On Thu, Nov 28, 2019 at 07:07:22PM +0800, Pengzhou Tang wrote:
> > Richard pointed out that he get incorrect results with the patch I
> > attached, there are bugs somewhere,
> > I fixed them now and attached the newest version, please refer
On Thu, Nov 28, 2019 at 07:07:22PM +0800, Pengzhou Tang wrote:
> Richard pointed out that he get incorrect results with the patch I
> attached, there are bugs somewhere,
> I fixed them now and attached the newest version, please refer to [1] for
> the fix.
Mr Robot is reporting that the latest pat
e/parallel_groupingsets_3>_3
>
> On Wed, Jul 31, 2019 at 4:07 PM Richard Guo wrote:
>
>> On Tue, Jul 30, 2019 at 11:05 PM Tomas Vondra <
>> tomas.von...@2ndquadrant.com> wrote:
>>
>>> On Tue, Jul 30, 2019 at 03:50:32PM +0800, Richard Guo wrote:
>>>
Richard Guo wrote:
>> >On Wed, Jun 12, 2019 at 10:58 AM Richard Guo wrote:
>> >
>> >> Hi all,
>> >>
>> >> Paul and I have been hacking recently to implement parallel grouping
>> >> sets, and here we have two implementations.
>&g
On Tue, Jul 30, 2019 at 11:05 PM Tomas Vondra
wrote:
> On Tue, Jul 30, 2019 at 03:50:32PM +0800, Richard Guo wrote:
> >On Wed, Jun 12, 2019 at 10:58 AM Richard Guo wrote:
> >
> >> Hi all,
> >>
> >> Paul and I have been hacking recently to implement par
On Tue, Jul 30, 2019 at 03:50:32PM +0800, Richard Guo wrote:
On Wed, Jun 12, 2019 at 10:58 AM Richard Guo wrote:
Hi all,
Paul and I have been hacking recently to implement parallel grouping
sets, and here we have two implementations.
Implementation 1
Attached is the patch
On Wed, Jun 12, 2019 at 10:58 AM Richard Guo wrote:
> Hi all,
>
> Paul and I have been hacking recently to implement parallel grouping
> sets, and here we have two implementations.
>
> Implementation 1
>
>
> Attached is the patch and also there is a
On Fri, Jun 14, 2019 at 12:02:52PM +1200, David Rowley wrote:
On Fri, 14 Jun 2019 at 11:45, Tomas Vondra wrote:
On Wed, Jun 12, 2019 at 10:58:44AM +0800, Richard Guo wrote:
># explain (costs off, verbose) select c1, c2, avg(c3) from t2 group by
>grouping sets((c1,c2), (c1));
>
On Fri, 14 Jun 2019 at 11:45, Tomas Vondra wrote:
>
> On Wed, Jun 12, 2019 at 10:58:44AM +0800, Richard Guo wrote:
> ># explain (costs off, verbose) select c1, c2, avg(c3) from t2 group by
> >grouping sets((c1,c2), (c1));
> > QUERY PLAN
> >
On Wed, Jun 12, 2019 at 10:58:44AM +0800, Richard Guo wrote:
Hi all,
Paul and I have been hacking recently to implement parallel grouping
sets, and here we have two implementations.
Implementation 1
Attached is the patch and also there is a github branch [1] for this
work
sult for each group.
> >
> > We are implementing parallel grouping sets in the same way. The only
> > difference is that in the final stage, the leader performs a grouping
> > sets aggregation, rather than a normal aggregation.
>
> Hi Richard,
>
> I think it was you an I
which
> that process is aware. Second, the partial results are transferred to
> the leader via the Gather node. Finally, the leader merges the partial
> results and produces the final result for each group.
>
> We are implementing parallel grouping sets in the same way. The only
> d
Hi all,
Paul and I have been hacking recently to implement parallel grouping
sets, and here we have two implementations.
Implementation 1
Attached is the patch and also there is a github branch [1] for this
work.
Parallel aggregation has already been supported in PostgreSQL
25 matches
Mail list logo
