Creating spatial indexes on geography slower than on geometry?

[Marco Boeringa]

Thanks for pointing that out Imre!

That recent work on parallel GiST builds looks very promising, and 
certainly the benchmarks are eyewatering. If this work makes it into 
PG18, it will be a game changer for spatially indexing OpenStreetMap 
Planet data, and make use of geography more realistic with ultra large 
datasets. The postprocessing on 'planet_osm_polygon' of Daylight v1.55 
took 24h33m with 'geography', while only 7h51m with 'geometry'. Both of 
these timings will be slashed to usable proportions with parallel build.

Marco

Op 2-11-2024 om 15:15 schreef Imre Samu:
>
>     Just wondering, given the fact that, as it now turns out, the
>     specific step to calculate the bounding boxes is such a large part
>     of the overall time of the building of the GiST spatial index,
>     shouldn't this particular step be parallelized in PostgreSQL via
>     parallel workers?
>
>
> related:  "WIP: parallel GiST index builds"  ( first mail: 2024-06-07 
> 17:41:10   last mail: 2024-10-31 18:05:43 )
> https://www.postgresql.org/message-id/flat/0e4f4af8-1088-4995-b2fb-8f92b5c6cef9%40enterprisedb.com
>
> Imre
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-users/attachments/20241102/7e18a8a7/attachment.htm>