[postgis-devel] Two geometries intersects but intersection is empty

[Pierre Racine]

+1 for JASPA... :-(

>-----Original Message-----
>From: postgis-devel-bounces at postgis.refractions.net [mailto:postgis-devel-
>bounces at postgis.refractions.net] On Behalf Of Paul Ramsey
>Sent: 3 juin 2010 15:22
>To: PostGIS Development Discussion
>Subject: Re: [postgis-devel] Two geometries intersects but intersection is empty
>
>Ah, yes, PostGIS is using its own code for the intersects test on
>point/polygon... hence the incompatibility. Hm hm.
>
>P
>
>On Thu, Jun 3, 2010 at 3:15 PM, Martin Davis <mbdavis at refractions.net> wrote:
>> Thanks.
>>
>> I was actually a bit puzzled by this, since in JTS the intersects and
>> intersection operations use exactly the same code to test whether the point
>> lies inside the polygon.  And in fact, when I run this case in JTS, it
>> reports that:
>>
>> intersects = FALSE
>> intersection = EMPTY
>>
>> which is consistent.
>>
>> It seems odd that PostGIS behaves differently.  It would be interesting to
>> try this directly in GEOS.  Although does PostGIS perhaps not use GEOS for
>> this particular case?
>> Also, the code run in JTS is of fairly recent vintage, so it's possible that
>> GEOS isn't reflecting the most up-to-date algorithm.
>>
>> Martin
>>
>> Pierre Racine wrote:
>>>
>>> SELECT st_intersects(geomA, geomB), astext(st_intersection(geomA, geomB))
>>> FROM (SELECT
>>>      st_geomfromtext('POINT(-75.5516150170498 49.2824791927083)',4326) AS
>>> geomA,
>>>
>>>
> geomfromewkb(decode('0103000020e610000001000000050000001c9f33ce69e352c0718f0e9a27a44840e77546654fe352
>c067d2ad7b35a448406cd4767448e352c0fc7fd3a900a44840a1fd63dd62e352c0073d34c8f2a348401c9f33ce69e352c0718f
>0e9a27a44840',
>>> 'hex')) AS geomB
>>> ) foo
>>>
>>>
>>>
>>>>
>>>> -----Original Message-----
>>>> From: postgis-devel-bounces at postgis.refractions.net
>>>> [mailto:postgis-devel-
>>>> bounces at postgis.refractions.net] On Behalf Of Paul Ramsey
>>>> Sent: 3 juin 2010 14:26
>>>> To: PostGIS Development Discussion
>>>> Subject: Re: [postgis-devel] Two geometries intersects but intersection
>>>> is empty
>>>>
>>>> FYI: select encode(asbinary(geom),'hex') from ...
>>>>
>>>> P
>>>>
>>>> On Thu, Jun 3, 2010 at 2:24 PM, Martin Davis <mbdavis at refractions.net>
>>>> wrote:
>>>>
>>>>>
>>>>> Pierre, can you provide the second geometry in WKBHex?  I can't read the
>>>>> string that you provided.
>>>>>
>>>>> Pierre Racine wrote:
>>>>>
>>>>>>
>>>>>> Here is the case:
>>>>>>
>>>>>> SELECT st_intersects(geomA, geomB), astext(st_intersection(geomA,
>>>>>> geomB))
>>>>>> FROM (SELECT
>>>>>>     st_geomfromtext('POINT(-75.5516150170498 49.2824791927083)',4326)
>>>>>> AS
>>>>>> geomA,
>>>>>>     geomfromewkb(E'\\001\\003\\000\\000
>>>>>>
>>>>>>
>>>>
>>>>
>>>>
>\\346\\020\\000\\000\\001\\000\\000\\000\\005\\000\\000\\000\\034\\2373\\316i\\343R\\300q\\217\\016\\2
>>>>
>>>>
>32\'\\244H@\\347uFeO\\343R\\300g\\322\\255{5\\244H at l\\324vtH\\343R\\300\\374\\177\\323\\251\\000\\244H
>>>>
>>>>
>@\\241\\375c\\335b\\343R\\300\\007=4\\310\\362\\243H@\\034\\2373\\316i\\343R\\300q\\217\\016\\232\'\\2
>>>> 44H@')
>>>>
>>>>>>
>>>>>> AS geomB
>>>>>> ) foo
>>>>>>
>>>>>> Sorry I can't provide it as wkt as the wkt version return the expected
>>>>>> result (and the wkb version does not (before I read Paul's last FAQ)).
>>>>>>
>>>>>> Pierre
>>>>>>
>>>>>>
>>>>>>
>>>>>>>
>>>>>>> -----Original Message-----
>>>>>>> From: postgis-devel-bounces at postgis.refractions.net
>>>>>>> [mailto:postgis-devel-
>>>>>>> bounces at postgis.refractions.net] On Behalf Of Martin Davis
>>>>>>> Sent: 3 juin 2010 12:15
>>>>>>> To: PostGIS Development Discussion
>>>>>>> Subject: Re: [postgis-devel] Two geometries intersects but
>>>>>>> intersection
>>>>>>> is empty
>>>>>>>
>>>>>>> Can you post the case?
>>>>>>>
>>>>>>> This can happen because the intersects predicate test is more accurate
>>>>>>> than the intersection computation.  (Technically speaking, it's
>>>>>>> possible
>>>>>>> to detect that two segments intersect, but be unable to compute an
>>>>>>> accurate point that represents the intersection).
>>>>>>>
>>>>>>> Pierre Racine wrote:
>>>>>>>
>>>>>>>
>>>>>>>>
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> Is it possible that st_intersects returns true on two geometries but
>>>>>>>> their st_intersection returns
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>> "GEOMETRYCOLLECTION EMPTY"?
>>>>>>>
>>>>>>>
>>>>>>>>
>>>>>>>> I have a nice case here...
>>>>>>>>
>>>>>>>> Pierre
>>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>>> postgis-devel mailing list
>>>>>>>> postgis-devel at postgis.refractions.net
>>>>>>>> http://postgis.refractions.net/mailman/listinfo/postgis-devel
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> Martin Davis
>>>>>>> Senior Technical Architect
>>>>>>> Refractions Research, Inc.
>>>>>>> (250) 383-3022
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> postgis-devel mailing list
>>>>>>> postgis-devel at postgis.refractions.net
>>>>>>> http://postgis.refractions.net/mailman/listinfo/postgis-devel
>>>>>>>
>>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> postgis-devel mailing list
>>>>>> postgis-devel at postgis.refractions.net
>>>>>> http://postgis.refractions.net/mailman/listinfo/postgis-devel
>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>> --
>>>>> Martin Davis
>>>>> Senior Technical Architect
>>>>> Refractions Research, Inc.
>>>>> (250) 383-3022
>>>>>
>>>>> _______________________________________________
>>>>> postgis-devel mailing list
>>>>> postgis-devel at postgis.refractions.net
>>>>> http://postgis.refractions.net/mailman/listinfo/postgis-devel
>>>>>
>>>>>
>>>>
>>>> _______________________________________________
>>>> postgis-devel mailing list
>>>> postgis-devel at postgis.refractions.net
>>>> http://postgis.refractions.net/mailman/listinfo/postgis-devel
>>>>
>>>
>>> _______________________________________________
>>> postgis-devel mailing list
>>> postgis-devel at postgis.refractions.net
>>> http://postgis.refractions.net/mailman/listinfo/postgis-devel
>>>
>>>
>>
>> --
>> Martin Davis
>> Senior Technical Architect
>> Refractions Research, Inc.
>> (250) 383-3022
>>
>> _______________________________________________
>> postgis-devel mailing list
>> postgis-devel at postgis.refractions.net
>> http://postgis.refractions.net/mailman/listinfo/postgis-devel
>>
>_______________________________________________
>postgis-devel mailing list
>postgis-devel at postgis.refractions.net
>http://postgis.refractions.net/mailman/listinfo/postgis-devel