Re: How to run in parallel in Postgres

[Lars Aksel Opsahl <Lars.Opsahl@xxxxxxxx>]

 >From: Ondrej Ivanič <ondrej.ivanic@gmail.com>

 >Sent: Saturday, December 7, 2019 2:23 AM

 >Cc: pgsql-performance@lists.postgresql.org <pgsql-performance@lists.postgresql.org>

 >Subject: Re: How to run in parallel in Postgres

 > 

 >Hi Lars,

 >

 >I have two suggestions:

 >

 >- `xargs` almost always present and it can run in parallel (-P) but script needs to be changed:

 >for((i=1;i<=28;i++)); do echo "SELECT find_overlap_gap_single_cell('test_data.overlap_gap_input_t1','geom',4258,'test_data.overlap_gap_input_t1_res',${I},28);"; done | xargs -n1 -P 10 psql ... 

 >

 >- `UNION ALL` might trigger parallel execution (you need to mess with the cost of the function and perhaps other settings):

 >SELECT

 > find_overlap_gap_single_cell('test_data.overlap_gap_input_t1','geom',4258,'test_data.overlap_gap_input_t1_res',1,28)

 >UNION ALL

 >SELECT

 > find_overlap_gap_single_cell('test_data.overlap_gap_input_t1','geom',4258,'test_data.overlap_gap_input_t1_res',2,28)

 >...

 >

 >

 >Cheers,

 >

Hi Ondrej

• Yes using xargs seems be an alternative to GNU parallel and I will have that in mind.
• I did a test using UNION ALL in the branch https://github.com/larsop/find-overlap-and-gap/tree/union_all_parallel but I was not able to trigger Postgres parallel .  That may be related to what Justin say about create tables.

Thanks.

Lars