[postgis-devel] GiST Sorting

[Paul Ramsey]

It does in fact come from gevel. Here's my fork, which is patched for Pg14

https://github.com/pramsey/gevel

P

> On Jan 12, 2022, at 4:09 PM, Giuseppe Broccolo <g.broccolo.7 at gmail.com> wrote:
> 
> Hi Paul,
> 
> I have a question: I see you used the function `gist_stat` to inspect the GiST indexes which is quite useful to check the overlaps between the boxes in the leaf nodes. I would like to use it as well,
> but unfortunately I wasn't able to find any reference beside the gevel repo. Could you tell me which extension you used to get this function?
> 
> Thanks in advance,
> Giuseppe.
> 
> Il giorno mar 11 gen 2022 alle ore 21:17 Paul Ramsey <pramsey at cleverelephant.ca> ha scritto:
> 
> 
> > On Jan 11, 2022, at 12:52 PM, Giuseppe Broccolo <g.broccolo.7 at gmail.com> wrote:
> > 
> > Hi Darafei,
> > 
> > On Sun, 9 Jan 2022, 14:12 Darafei "Komяpa" Praliaskouski, <me at komzpa.net> wrote:
> > Hi everyone,
> > 
> > After several weeks of research, testing and experiments we believe we found a solution that does not cause the select performance degradation (or sometimes brings profit) and still builds the index faster using the sorting build method. 
> > 
> > The patch is posted on pgsql-hackers by Aliaskandr Kalenik in this thread:
> > 
> > https://www.postgresql.org/message-id/CAHqSB9jqtS94e9%3D0vxqQX5dxQA89N95UKyz-%3DA7Y%2B_YJt%2BVW5A%40mail.gmail.com
> > 
> > Patch on commitfest:
> > 
> > https://commitfest.postgresql.org/36/3487/
> > 
> > Sergei Shoulbakov implemented the benchmark to prove the effect. Results and notes are posted here:
> > https://www.postgresql.org/message-id/3aa6ba30-e9d8-10ef-753f-8deea5f196d0%40kontur.io
> > 
> > We need your help to get it reviewed and merged to get the benefits for everyone in PG15. 
> > 
> > That's cool! What can we do to help?
> 
> Well, one thing we definitely need is to compare the performance of the patched sortsupport method to the vanilla index build, across a variety of inputs and workloads. My example data was interesting, as a heterogenous set of data (differing object spatial sizes, differing densities across the data set) but there's lot of other different data that should be benchmarked. Get your favourite data sets and see what happens. Tests on small tables might also have some quirks that don't show up in large table tests. Harder to measure, but worth measuring.
> 
> ATB,
> P
> 
> 
> > 
> > Giuseppe. 
> > _______________________________________________
> > postgis-devel mailing list
> > postgis-devel at lists.osgeo.org
> > https://lists.osgeo.org/mailman/listinfo/postgis-devel
> 
> _______________________________________________
> postgis-devel mailing list
> postgis-devel at lists.osgeo.org
> https://lists.osgeo.org/mailman/listinfo/postgis-devel
> _______________________________________________
> postgis-devel mailing list
> postgis-devel at lists.osgeo.org
> https://lists.osgeo.org/mailman/listinfo/postgis-devel