Re: [HACKERS] Parallel Aggregation support for aggregate functions that use transitions not implemented for array_agg

2017-06-18 Thread Tomas Vondra
Hi, On 6/7/17 5:52 AM, Regina Obe wrote: On 6/6/17 13:52, Regina Obe wrote: It seems CREATE AGGREGATE was expanded in 9.6 to support parallelization of aggregate functions using transitions, with the addition of serialfunc and deserialfunc to the aggregate definitions.

Re: [HACKERS] Parallel Aggregation support for aggregate functions that use transitions not implemented for array_agg

2017-06-08 Thread Ashutosh Bapat
On Wed, Jun 7, 2017 at 10:55 PM, Robert Haas wrote: > On Tue, Jun 6, 2017 at 3:23 PM, David Fetter wrote: >> I'd bet on lack of tuits. > > I expect that was part of it. Another thing to consider is that, for > numeric aggregates, the transition values

Re: [HACKERS] Parallel Aggregation support for aggregate functions that use transitions not implemented for array_agg

2017-06-07 Thread Robert Haas
On Tue, Jun 6, 2017 at 3:23 PM, David Fetter wrote: > I'd bet on lack of tuits. I expect that was part of it. Another thing to consider is that, for numeric aggregates, the transition values don't generally get larger as you aggregate, but for something like string_agg(), they

Re: [HACKERS] Parallel Aggregation support for aggregate functions that use transitions not implemented for array_agg

2017-06-06 Thread Andres Freund
On 2017-06-07 00:03:15 -0400, Tom Lane wrote: > Andres Freund writes: > > On 2017-06-06 23:32:53 -0400, Peter Eisentraut wrote: > >> I'm not sure how you would parallelize these, since in most uses you > >> want to have a deterministic output order. > > > Unless you specify

Re: [HACKERS] Parallel Aggregation support for aggregate functions that use transitions not implemented for array_agg

2017-06-06 Thread Tom Lane
Andres Freund writes: > On 2017-06-06 23:32:53 -0400, Peter Eisentraut wrote: >> I'm not sure how you would parallelize these, since in most uses you >> want to have a deterministic output order. > Unless you specify ORDER BY you don't really have that anyway, consider >

Re: [HACKERS] Parallel Aggregation support for aggregate functions that use transitions not implemented for array_agg

2017-06-06 Thread Regina Obe
> On 6/6/17 13:52, Regina Obe wrote: >> It seems CREATE AGGREGATE was expanded in 9.6 to support >> parallelization of aggregate functions using transitions, with the >> addition of serialfunc and deserialfunc to the aggregate definitions. >> >>

Re: [HACKERS] Parallel Aggregation support for aggregate functions that use transitions not implemented for array_agg

2017-06-06 Thread Andres Freund
On 2017-06-06 23:32:53 -0400, Peter Eisentraut wrote: > On 6/6/17 13:52, Regina Obe wrote: > > It seems CREATE AGGREGATE was expanded in 9.6 to support parallelization of > > aggregate functions using transitions, with the addition of serialfunc and > > deserialfunc to the aggregate definitions.

Re: [HACKERS] Parallel Aggregation support for aggregate functions that use transitions not implemented for array_agg

2017-06-06 Thread Peter Eisentraut
On 6/6/17 13:52, Regina Obe wrote: > It seems CREATE AGGREGATE was expanded in 9.6 to support parallelization of > aggregate functions using transitions, with the addition of serialfunc and > deserialfunc to the aggregate definitions. > >

Re: [HACKERS] Parallel Aggregation support for aggregate functions that use transitions not implemented for array_agg

2017-06-06 Thread Andres Freund
On 2017-06-06 12:23:49 -0700, David Fetter wrote: > On Tue, Jun 06, 2017 at 01:52:45PM -0400, Regina Obe wrote: > > It seems CREATE AGGREGATE was expanded in 9.6 to support parallelization of > > aggregate functions using transitions, with the addition of serialfunc and > > deserialfunc to the

Re: [HACKERS] Parallel Aggregation support for aggregate functions that use transitions not implemented for array_agg

2017-06-06 Thread David Fetter
On Tue, Jun 06, 2017 at 01:52:45PM -0400, Regina Obe wrote: > It seems CREATE AGGREGATE was expanded in 9.6 to support parallelization of > aggregate functions using transitions, with the addition of serialfunc and > deserialfunc to the aggregate definitions. > >