[postgis-users] Looking for help figuring out overlays

Marcin Mionskowski mionskowskimarcin at gmail.com
Tue Sep 1 00:22:56 PDT 2020


Sorry for the late answer.
This problem is probably related to ST_Union erroneous behavior in some
cases. It was described on this list last month (with multilinestring
example). You can try ST_MemUnion - it's slower but should yield correct
results.
Greetings,
Marcin

sob., 13 cze 2020 o 03:54 Jonathan Schultz <jonathan at schultz.la> napisaƂ(a):

> Hello,
>
> I think I've reached the limit of my PostGIS skills so hoping someone can
> help me solve a problem.
>
> What I'm trying to do is create overlays from a collection of overlapping
> polygons as described here:
> http://blog.cleverelephant.ca/2019/07/postgis-overlays.html
>
> The data I am working from is a publicly available dataset of the fire
> history of Western Australia:
> https://catalogue.data.wa.gov.au/dataset/dbca-fire-history I downloaded
> the data as CSV and loaded it into a PostgreSQL table. That much seems fine.
>
> I then select a region and dump the MULTIPOLYGONs into POLYGONs using the
> following query:
>
> > CREATE TABLE dbca_fire_history_dump AS
> >     SELECT fih_perimeter, fih_author, fih_season1, fih_capt_meth,
> fih_year1, fih_district, fih_burn_purp, fih_detect_fdi, fih_poly_type,
> fih_fire_type, fih_master_key, fih_number, fih_date1, fih_cause,
> fih_ignit_type, fih_comment, fih_hectares, fih_hist_distr, fih_fire_seaso,
> fih_name, (st_dump(geometry)).geom AS geometry FROM dbca_fire_history
> >     WHERE st_intersects(geometry, st_makeenvelope(<longmin>, <latmin>,
> <longmax>, <latmax>, 4326));
>
> and from there create the boundary polygon using this query:
>
> > CREATE TABLE dbca_fire_history_boundaries AS
> >   SELECT ST_Union(ST_ExteriorRing(geometry)) AS geometry
> >     FROM dbca_fire_history_dump;
>
> This is where things get weird. If I select a small enough region (eg 36S
> 116E to 35S 117E) the result looks fine. But if I select a slightly larger
> region (eg 36S 116E to 34S 117E) then the result is very strange - the
> geometry seems to have snapped to a much larger grid to the extent that it
> bears no obvious resemblance to the data.
>
> I've pasted images that demonstrate the result here:
> https://pasteboard.co/JcPeDUk.png and here:
> https://pasteboard.co/JcPfju0H.png
>
> Thanks in advance for any help or suggestions.
>
> Cheers,
> Jonathan
> _______________________________________________
> postgis-users mailing list
> postgis-users at lists.osgeo.org
> https://lists.osgeo.org/mailman/listinfo/postgis-users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-users/attachments/20200901/14ccd26f/attachment.html>


More information about the postgis-users mailing list