[postgis-users] very strange bug generating incorrect ST_Distance calculations

[David M. Kaplan]

On 10/01/2019 12:26, David M. Kaplan wrote:
> Comparing the results of _ST_DistanceTree and _ST_DistanceUnCached is
> a good way to tease out whether there are differences between the
> implementations without having to fight the caching machinery along
> the way.
>
> P.

Thanks for the responses. I tried with _ST_DistanceTree and 
_ST_DistanceUnCached
and as predicted, tree gives 0 and uncached gives the correct answer. 
What is the appropriate work around to avoid this issue? Who should I 
report this problem to? The maintainers of libgeos?

Also, how should I interpret that this error only occurs when one of the 
geometries is a multipolygon?

In response to Raúl, I found that both queries give 0 for postgis 2.5.0 
r16836, but only the multiline query gives zero for POSTGIS="2.4.3 
r16312"...

Cheers,
David

>
> On Thu, Jan 10, 2019 at 7:00 AM Raúl Marín Rodríguez
> <rmrodriguez at carto.com  <https://lists.osgeo.org/mailman/listinfo/postgis-users>> wrote:
> >//>/Hi, />//>/In my case I get dist_geography between gid 1 and gid 3 0 in both 
> queries. />//>/This looks like a possible issue around the geos cache, something 
> similar />/to https://trac.osgeo.org/postgis/ticket/4269. Can you run them with />/EXPLAIN ANALYZE? />//>/-- />/Raúl Marín Rodríguez />/carto.com />/_______________________________________________ />/postgis-users mailing list />/postgis-users at lists.osgeo.org 
> <https://lists.osgeo.org/mailman/listinfo/postgis-users> />/https://lists.osgeo.org/mailman/listinfo/postgis-users /

> Hi,
>
> I have found what seems to be an extremely bizarre bug. I am not sure 
> if it is a postgis issue or postgresql issue, but it leads 
> ST_Distance(geography(Polygon,4326),geography(MultiPolygon,4326)) to 
> return 0 (zero) when the true answer is not zero under certain 
> peculiar situations.
>
> The table data for a set of polygons that generate the error can be 
> downloaded here:
>
> http://www.davidmkaplan.fr/bad_geography_dist_calcs.csv
>
> Once you have that data, the code to generate the error is:
>
> CREATE TEMP TABLE tt (gid int PRIMARY KEY, geom geometry);
>
> \copy tt FROM 'bad_geography_dist_calcs.csv' WITH (FORMAT 'csv',HEADER 
> TRUE)
>
> -- Incorrect dist_geography for t1.gid=1 and t2.gid=3
> SELECT t1.gid AS gid1, t2.gid AS gid2,
>        ST_Distance(t1.geom,t2.geom) AS dist_lonlat,
>        ST_Distance(ST_Transform(t1.geom,26918),
>                    ST_Transform(t2.geom,26918)) AS dist_utm,
>        ST_Distance(t1.geom::geography,t2.geom::geography) AS 
> dist_geography,
> ST_Distance((ST_Dump(t1.geom)).geom::geography,(ST_Dump(t2.geom)).geom::geography) 
> AS dist_geography_dump
> FROM tt t1 JOIN tt t2 ON t1.gid<t2.gid
> WHERE t1.gid=1
> ;
>
> -- No error just by specifying t2.gid=3 in the WHERE clause
> SELECT t1.gid AS gid1, t2.gid AS gid2,
>        ST_Distance(t1.geom,t2.geom) AS dist_lonlat,
>        ST_Distance(ST_Transform(t1.geom,26918),
>                    ST_Transform(t2.geom,26918)) AS dist_utm,
>        ST_Distance(t1.geom::geography,t2.geom::geography) AS 
> dist_geography,
> ST_Distance((ST_Dump(t1.geom)).geom::geography,(ST_Dump(t2.geom)).geom::geography) 
> AS dist_geography_dump
> FROM tt t1 JOIN tt t2 ON t1.gid<t2.gid
> WHERE t1.gid=1 AND t2.gid=3
> ;
>
> Hopefully, if I am not crazy, you will have a 0 value of 
> dist_geography in the first query for t2.gid=3, but not in the second.
>
> Does anyone have any idea what could be causing this? To be honest, I 
> am not really certain that I know where to start do diagnose the 
> problem...
>
> The specifics of my installation are:
>
> POSTGIS="2.4.3 r16312" PGSQL="100" GEOS="3.7.0-CAPI-1.11.0 673b9939" 
> PROJ="Rel. 4.9.3, 15 August 2016" GDAL="GDAL 2.3.2, released 
> 2018/09/21" LIBXML="2.9.4" LIBJSON="0.12.1" LIBPROTOBUF="1.2.1" RASTER
>
> psql 10.6 (Ubuntu 10.6-0ubuntu0.18.04.1))
>
> Ubuntu 18.04.1 LTS
>
> I have also tested this with postgis 2.5.0 r16836 and there the 
> incorrect 0 is in both queries!?!?!?
>
> Thanks,
> David
>
>

-- 
**********************************
David M. Kaplan
Charge de Recherche 1

Institut de Recherche pour le Developpement (IRD)
UMR MARBEC (IRD/Ifremer/CNRS/UMII)
av. Jean Monnet
CS 30171
34203 Sete cedex
France

Email: david.kaplan at ird.fr
Phone: +33 (0)4 99 57 32 25
Fax: +33 (0)4 99 57 32 95

http://www.umr-marbec.fr/kaplan-david.html
http://www.davidmkaplan.fr/
**********************************

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-users/attachments/20190110/85ef231b/attachment.html>