[postgis-devel] ST_Union Parallel Experiment

[Regina Obe]

>  Not marking something as PARALLEL SAFE just plainly disables parallel for everything in the query, this was the main reason for PARALLEL SAFE. Without Paul's modification you can still get

 

> benefit from parallelism, in form of say  

 

> SELECT class, ST_Union(ST_Buffer(geom, 1)) from sometable group by class;

I thought not marking it as higher than PARALLEL Restricted would do that.  So PARALLEL restricted would do the same, but PARALLEL SAFE means it can run inside a PARALLEL worker.

 

What I was asking was:

 

SELECT class, ST_Union(geom)

   FROM sometable GROUP BY class

 

 

Someone claimed it was possible for the sorting to happen in parallel, such that all the geometries that fall in a group would be guaranteed to be in one worker.

Then as you said there would be no need to call a combinefunc since all of class  A – would have been taken care of and combinefunc would never need to be called?

 

Beautiful diagram by the way.  But perhaps, you need a combinefunc in there anyway even if it's never called because the parallelizer can't guarantee it can allocate all the items that belong in one group to the same worker.

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-devel/attachments/20180313/9683c24d/attachment.html>