[postgis-users] Use of tolerances in topology

Jose Carlos Martinez jomarlla at cgf.upv.es
Tue Apr 17 09:29:47 PDT 2012


Thanks Sandro, I understand now.

About  trying to enhance ST_Dwithin, last time I checked PostGIS code it 
was implemented using brute force (after comparing the boxes). Maybe one 
way to improve the algorithm (DT_Dwithin) will be to use spatial indexes 
inside a geometry. According to my tests (long time ago with JTS) I 
remember it could be faster to build a spatial index every time for a 
geometry and use it than using brute force (with geometries with more 
than 20 vertexes or so).

Anyways Im not sure if PostGIS is still using brute force with 
st_dwithin, if so then forget everything I said.
Regards,

On 17/04/2012 17:46, Sandro Santilli wrote:
> On Tue, Apr 17, 2012 at 05:35:42PM +0200, Jose Carlos Martinez wrote:
>> Hi Sandro, Im learning from topology.sql how the topology functions
>> are working with more detail.
>>
>> I have a question about how the tolerances are managed, for example:
>>
>> I see (ST_AddIsoEdge) that you have used some times spatial
>> predicates as st_contains or st_intersects. As you know these
>> predicates (and others too) just work in a good way if the
>> geometries are snapped to each other previously,  so my question is
>> why you didnt use st_Dwithin with a tolerance? or maybe these
>> functions expect the geometries to be already snapped by totopogeom
>> for example and they shouldnt be used directly?
> Those functions are ISO functions and as such are not specified to
> deal with a tolerance. Handling of tolerance is not always
> straightforward as using ST_DWithin. It may also have a big cost.
>
> So take all those functions as being expected to be already snapped.
>
> I'm open to suggestion about enhanced versions taking tolerance into
> consideration. But it must be possible to invoke such functions in
> a fast way, when caller already did their checking and noding and what
> not.
>
> The idea is that all the ST_ prefixed functions end up being just wrappers
> around more flexible functions in core.
>
> --strk;
>
>    ,------o-.
>    |   __/  |    Delivering high quality PostGIS 2.0 !
>    |  / 2.0 |    http://strk.keybit.net - http://vizzuality.com
>    `-o------'
>
> _______________________________________________
> postgis-users mailing list
> postgis-users at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-users




More information about the postgis-users mailing list