[postgis-devel] Which intersection is faster?

[Jorge Arévalo]

That's a good tip. I'll take a deeper look on GDAL cache (not the same concept, but it could help) and try to implement this idea.  

Many thanks!   

--  
Jorge Arevalo
Freelance developer

http://www.krop.com/jorgearevalo
http://about.me/jorgeas80

Enviado con Sparrow (http://www.sparrowmailapp.com/?sig)


El sábado 23 de marzo de 2013 a las 19:42, Bborie Park escribió:

> Your best bet would be to convert the geometry to an aligned raster
> once, cache that raster and then run the intersection test.
> Otherwise, things get converted needlessly and repeatedly.
>  
> -bborie
>  
> On Sat, Mar 23, 2013 at 11:12 AM, Jorge Arévalo
> <jorgearevalo at libregis.org (mailto:jorgearevalo at libregis.org)> wrote:
> > Ok, understood.
> >  
> > The point is I'm trying to make the GDAL PostGIS Raster driver faster, and the main bottleneck is in the intersection between raster columns and user requested window (I construct a polygon from the user's request coords and instersect that polygon with the raster data).
> >  
> > Swapping raster - vector objects was my first (naive) approach. Any (for sure smarter) suggestions are welcome.
> >  
> > --
> > Jorge Arevalo
> > Freelance developer
> >  
> > http://www.krop.com/jorgearevalo
> > http://about.me/jorgeas80
> >  
> > Enviado con Sparrow (http://www.sparrowmailapp.com/?sig)
> >  
> >  
> > El sábado 23 de marzo de 2013 a las 18:53, Bborie Park escribió:
> >  
> > > There's not too much difference between 2.0 and 2.1. I think the
> > > major change is ST_Intersects(raster, geometry) =
> > > ST_Intersects(geometry, raster). Basically, if one of the arguments
> > > is a geometry, the op is done in geometry space (raster is converted
> > > to a set of geometries).
> > >  
> > > -bborie
> > >  
> > > On Sat, Mar 23, 2013 at 10:50 AM, Jorge Arévalo
> > > <jorgearevalo at libregis.org (mailto:jorgearevalo at libregis.org)> wrote:
> > > > Many thanks for the quick response. I can see is pretty fast. But, what about previous version of PostGIS? Say 2.0. Is there any significant difference between both versions?
> > > >  
> > > > --
> > > > Jorge Arevalo
> > > > Freelance developer
> > > >  
> > > > http://www.krop.com/jorgearevalo
> > > > http://about.me/jorgeas80
> > > >  
> > > > Enviado con Sparrow (http://www.sparrowmailapp.com/?sig)
> > > >  
> > > >  
> > > > El sábado 23 de marzo de 2013 a las 18:30, Bborie Park escribió:
> > > >  
> > > > > I'd suggest using the variant that runs in raster space. If I
> > > > > remember correctly, some changes have been made to this in 2.1.
> > > > >  
> > > > > -bborie
> > > > >  
> > > > > On Sat, Mar 23, 2013 at 10:20 AM, Jorge Arévalo
> > > > > <jorgearevalo at libregis.org (mailto:jorgearevalo at libregis.org)> wrote:
> > > > > > Hello,
> > > > > >  
> > > > > > I don't know if this is a naive question, but in ST_Intersects description (http://postgis.net/docs/manual-2.0/RT_ST_Intersects.html) I read:
> > > > > >  
> > > > > > "Depending on the order that the raster and geometry is passed to ST_Intersects(), the test will operate in either raster-space or geometry-space. If ST_Intersects(raster, ....), the test is in raster-space (the geometry is converted to a raster). If ST_Intersects(geometry, ...), the test is in geometry-space (the raster is converted to a set of pixel polygons)."
> > > > > >  
> > > > > > My question: in equal conditions, which version should be faster?
> > > > > >  
> > > > > > ST_Intersects(raster, vector) --> in raster space
> > > > > > ST_Intersects(vector, raster) --> in vector space
> > > > > >  
> > > > > > Many thanks in advance,
> > > > > >  
> > > > > > --
> > > > > > Jorge Arevalo
> > > > > > Freelance developer
> > > > > >  
> > > > > > http://www.krop.com/jorgearevalo
> > > > > > http://about.me/jorgeas80
> > > > > >  
> > > > > > Enviado con Sparrow (http://www.sparrowmailapp.com/?sig)
> > > > > >  
> > > > > >  
> > > > > > _______________________________________________
> > > > > > postgis-devel mailing list
> > > > > > postgis-devel at lists.osgeo.org (mailto:postgis-devel at lists.osgeo.org)
> > > > > > http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-devel
> > > > >  
> > > > >  
> > > > >  
> > > > >  
> > > > >  
> > > > >  
> > > > > _______________________________________________
> > > > > postgis-devel mailing list
> > > > > postgis-devel at lists.osgeo.org (mailto:postgis-devel at lists.osgeo.org)
> > > > > http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-devel
> > > >  
> > > >  
> > > >  
> > > >  
> > > >  
> > > >  
> > > >  
> > > > _______________________________________________
> > > > postgis-devel mailing list
> > > > postgis-devel at lists.osgeo.org (mailto:postgis-devel at lists.osgeo.org)
> > > > http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-devel
> > >  
> > >  
> > >  
> > >  
> > > _______________________________________________
> > > postgis-devel mailing list
> > > postgis-devel at lists.osgeo.org (mailto:postgis-devel at lists.osgeo.org)
> > > http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-devel
> >  
> >  
> >  
> >  
> >  
> > _______________________________________________
> > postgis-devel mailing list
> > postgis-devel at lists.osgeo.org (mailto:postgis-devel at lists.osgeo.org)
> > http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-devel
>  
>  
> _______________________________________________
> postgis-devel mailing list
> postgis-devel at lists.osgeo.org (mailto:postgis-devel at lists.osgeo.org)
> http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-devel