How to launch parallel aggregations ?

From: Esteban Zimanyi <ezimanyi(at)ulb(dot)ac(dot)be>
To: pgsql-hackers(at)lists(dot)postgresql(dot)org
Subject: How to launch parallel aggregations ?
Date: 2021-05-18 09:04:51
Message-ID: CAPqRbE5Mvc_ed7TmrOnE_VmnEzFDgSWiHctosQLotZyJGeSNJw@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Dear all

In MobilityDB we have defined parallel aggregations with a combine
function, e.g.,

CREATE AGGREGATE extent(tbox) (
SFUNC = tbox_extent_transfn,
STYPE = tbox,
COMBINEFUNC = tbox_extent_combinefn,
PARALLEL = safe
);

We would like to trigger the combine functions in the coverage tests but
for this it is required that the tables are VERY big. In particular for the
above aggregation, the combine function only is triggered when the table
has more than 300K rows.

As it is not very effective to have such a big table in the test database
used for the regression and the coverage tests I wonder whether it is
possible to set some parameters to launch the combine functions with tables
of, e.g., 10K rows, which are the bigger tables in our regression test
database.

Many thanks for your insights !

Esteban

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Greg Nancarrow 2021-05-18 09:14:37 Re: Re: Parallel scan with SubTransGetTopmostTransaction assert coredump
Previous Message Magnus Hagander 2021-05-18 08:54:37 Re: allow specifying direct role membership in pg_hba.conf