[geos-devel] Should DistanceWithin(EMPTY, ANYTHING, ANYDISTANCE) always return false ?

[Paul Ramsey]

Doesn't seem super important to back-port. To the extent that it's extant behaviour, I think we should just Let It Be.

P

> On Oct 4, 2021, at 3:43 PM, Sandro Santilli <strk at kbt.io> wrote:
> 
> I pushed a commit to main branch always returning false.
> 
> Same on the PostGIS side, but I don't know up to which version I
> should backport that. If you have old versions around, could you
> please leave a comment in https://trac.osgeo.org/postgis/ticket/5008 ?
> 
> --strk;
> 
> On Mon, Oct 04, 2021 at 07:56:34PM +0300, Darafei "Komяpa" Praliaskouski wrote:
>> Points on infinite distance are still on the extended R2+ plane. The empty
>> points aren't.
>> 
>> This one should be "always false", and make sure that "POINT(INF INF)" is
>> at INF distance from anything (can be true). POINT(Inf Inf) != POINT(NaN
>> NaN) in this aspect.
>> 
>> On Mon, Oct 4, 2021 at 6:50 PM Paul Ramsey <pramsey at cleverelephant.ca>
>> wrote:
>> 
>>> https://trac.osgeo.org/postgis/wiki/DevWikiEmptyGeometry
>>> 
>>> 
>>> pramsey=# select st_distance('POINT(1 1)','POINT EMPTY');
>>> st_distance
>>> -------------
>>> 
>>> (1 row)
>>> 
>>> Interesting note in the wiki about returning NULL for DWithin.
>>> 
>>> Current behaviour.
>>> 
>>> pramsey=# select st_dwithin('POINT(1 1)','POINT EMPTY', 1);
>>> st_dwithin
>>> ------------
>>> f
>>> (1 row)
>>> 
>>> pramsey=# select st_dwithin('POINT(1 1)','POINT EMPTY', 'Inf'::float8);
>>> st_dwithin
>>> ------------
>>> t
>>> (1 row)
>>> 
>>> I could see going for "always false" since we're almost there for non-inf
>>> values.
>>> 
>>> P
>>> 
>>>> On Oct 4, 2021, at 8:08 AM, Sandro Santilli <strk at kbt.io> wrote:
>>>> 
>>>> On Mon, Oct 04, 2021 at 05:02:24PM +0200, Sandro Santilli wrote:
>>>>> I'd think yes, honestly, but the current code returns TRUE sometimes,
>>>>> particularly for MAXDISTANCE being Infinite...
>>>>> 
>>>>> As I think GEOS is not being used by phylosophers, I can't find any
>>>>> reason why would anyone want a TRUE result from a DistanceWithin()
>>>>> query involving an EMPTY geometry.
>>>> 
>>>> We have the same issue on PostGIS:
>>>> https://trac.osgeo.org/postgis/ticket/5008
>>>> 
>>>> --strk;
>>>> _______________________________________________
>>>> geos-devel mailing list
>>>> geos-devel at lists.osgeo.org
>>>> https://lists.osgeo.org/mailman/listinfo/geos-devel
>>> 
>>> _______________________________________________
>>> geos-devel mailing list
>>> geos-devel at lists.osgeo.org
>>> https://lists.osgeo.org/mailman/listinfo/geos-devel
>>> 
> 
>> _______________________________________________
>> geos-devel mailing list
>> geos-devel at lists.osgeo.org
>> https://lists.osgeo.org/mailman/listinfo/geos-devel
> 
> 
> -- 
> 
>  Libre GIS consultant/developer
>  https://strk.kbt.io/services.html
> _______________________________________________
> geos-devel mailing list
> geos-devel at lists.osgeo.org
> https://lists.osgeo.org/mailman/listinfo/geos-devel