[postgis-users] diffrent execution plan on Postgres 9.5 and Postgres 11 for ST_union and performance problem Postgres 11

[Lars Aksel Opsahl]

Hi

Thanks a lot for your help everybody.

Lars

________________________________
From: postgis-users <postgis-users-bounces at lists.osgeo.org> on behalf of Martin Davis <mtnclimb at gmail.com>
Sent: Thursday, April 11, 2019 10:03 PM
To: PostGIS Users Discussion
Subject: Re: [postgis-users] diffrent execution plan on Postgres 9.5 and Postgres 11 for ST_union and performance problem Postgres 11

Well, we have figured out what the problem is - it's a known regression in the GEOS UnaryUnion code, reported in this ticket [1]

Unfortunately there's no fix for this yet, but we'll try and escalate this.

[1] https://trac.osgeo.org/geos/ticket/867

On Thu, Apr 11, 2019 at 11:07 AM Martin Davis <mtnclimb at gmail.com<mailto:mtnclimb at gmail.com>> wrote:


On Thu, Apr 11, 2019 at 3:27 AM Lars Aksel Opsahl <Lars.Opsahl at nibio.no<mailto:Lars.Opsahl at nibio.no>> wrote:
Can it be problem related ST_Union and aggregate ?

Can you test the performance of aggregate ST_Union on it's own by dropping the GROUP BY and instead using a WHERE clause to union a small-ish subset of the geometries?  (But large enough to give a measurable query time.)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-users/attachments/20190411/966b27ec/attachment.html>