On Wed, Mar 16, 2011 at 2:53 PM, Pierre Racine <[email protected]> wrote: > Yes, this is ticket 837 > > http://trac.osgeo.org/postgis/ticket/837 > > Jorge, are those numerous warning normal?
The warnings are normal if debug option is enabled. But what strikes me is printf shows real and truncated value as same value. I think this is caused because both values are compared in this way: if (doubleValue1 != doubleValue2) printf(....) instead of if (fabs(doubleValue1 - doubleValue2) > FLT_EPSILON) printf(....) So, this warning message is probably generated in setBits function. Anyway, I don't think that's the reason of the memory error. -- Jorge Arévalo Internet & Mobilty Division, DEIMOS [email protected] http://es.linkedin.com/in/jorgearevalo80 http://mobility.grupodeimos.com/ http://gis4free.wordpress.com http://geohash.org/ezjqgrgzz0g > ________________________________________ > From: [email protected] > [[email protected]] On Behalf Of Andreas Forø > Tollefsen [[email protected]] > Sent: Wednesday, March 16, 2011 9:24 AM > To: PostGIS Users Discussion > Subject: Re: [postgis-users] PostGIS Raster Decimal? > > Is the above memory issue known for the ST_MapAlgebra function? > Did you experience this Regina? > > I will try with some other rasters. > The WARNING lines are just a sample. I think it went through a couple of > thousands before it ran out of memory. > Is it not strange that the truncated values are the same as the original > values? > > Andreas > > 2011/3/15 Andreas Forø Tollefsen > <[email protected]<mailto:[email protected]>> > Hi Pierre, > I got memory issues as well. > I tried with this query: > SELECT ST_MapAlgebra(rast, 'rast*10') INTO algtest FROM gpw90; > > > WARNING: Pixel value for 32BF band got truncated from 47519.1 to 47519.1 > WARNING: Pixel value for 32BF band got truncated from 185494 to 185494 > WARNING: Pixel value for 32BF band got truncated from 28485.1 to 28485.1 > WARNING: Pixel value for 32BF band got truncated from 10927.1 to 10927.1 > > > ERROR: out of memory > DETAIL: Failed on request of size 19. > > ********** Error ********** > > ERROR: out of memory > SQL state: 53200 > Detail: Failed on request of size 19. > > 2011/3/14 Paragon Corporation <[email protected]<mailto:[email protected]>> > Pierre, > > Actually the single raster version is documented. Though I did run into > memory issues -- so still think we need some work on that. > > http://www.postgis.org/documentation/manual-svn/RT_ST_MapAlgebra.html > > Regina > http://www.postgis.us > > > > ________________________________ > From: > [email protected]<mailto:[email protected]> > > [mailto:[email protected]<mailto:[email protected]>] > On Behalf Of Andreas Forø Tollefsen > Sent: Monday, March 14, 2011 2:03 PM > To: PostGIS Users Discussion > Cc: Pierre Racine > > Subject: Re: [postgis-users] PostGIS Raster Decimal? > > Hi Pierre, > > I will check it out tomorrow and give some feedback. > > Cheers, > Andreas > > 2011/3/14 Pierre Racine > <[email protected]<mailto:[email protected]>> > I’m very aware of the problem. Will try to fix it the best we can. > > Why not testing ST_MapAlgebra and give us your feedback? It is still to > document but the specifications are in this page: > > http://trac.osgeo.org/postgis/wiki/WKTRaster/SpecificationWorking02 > > Pierre > > From: > [email protected]<mailto:[email protected]> > > [mailto:[email protected]<mailto:[email protected]>] > On Behalf Of Andreas Forø Tollefsen > Sent: 14 mars 2011 11:24 > To: PostGIS Users Discussion > Subject: Re: [postgis-users] PostGIS Raster Decimal? > > Hi Pierre. > > Yes. I think that would be an appropriate temporary fix. > Still, a bit problematic when working with climate data which very often > include decimal digits and often have weekly and monthly rasters for 100's of > years :) > > Good luck. > > Best, > Andreas > > 2011/3/14 Pierre Racine > <[email protected]<mailto:[email protected]>> > Andreas, > > This is a well known problem. See ticket > http://trac.osgeo.org/postgis/ticket/650 > > I will discuss this problem with Frank this week at the code sprint in > Montreal. > > For now you can use some GIS with map algebra capability to multiply your > raster by 10 before importing or test our brand new ST_MapAlgebra function > (still to debug). > > Pierre > >>-----Original Message----- >>From: Pierre Racine >>Sent: 14 mars 2011 10:55 >>To: Pierre Racine >>Subject: FW: [postgis-users] PostGIS Raster Decimal? >> >> >> >>From: >>[email protected]<mailto:[email protected]> >> [mailto:postgis-users-<mailto:postgis-users-> >>[email protected]<mailto:[email protected]>] On >>Behalf Of Andreas Forø Tollefsen >>Sent: 14 mars 2011 10:53 >>To: PostGIS Users Discussion >>Subject: [postgis-users] PostGIS Raster Decimal? >> >>Hi. Again. >> >>I have been working on some raster data in PostGIS lately. >>However, i have some issues with the raster values become integer after >>importing the sql to postgis. >>For instance this procedure: >> >>C:\prio_grid\source\gpw>c:\python26\python >>c:\prio_grid\script\raster2pgsql.py - >>r c:\prio_grid\source\gpw\lrc30p90\glp90ag30\w001001.adf -t gpw90 -s 4326 -o >>gpw >>90.sql -I -M >> >>Then to database: >>C:\prio_grid\source\gpw>psql -h 192.168.1.55 -d priogrid -f gpw90.sql >> >>Querying this data: >>SELECT gpw90.gid, ((gpw90.gpw90val).val) AS gpw90 >>INTO popgrid >>FROM (SELECT priogrid_land.gid, ST_Intersection(gpw90.rast, >>priogrid_land.centroid) AS gpw90val FROM >>gpw90, priogrid_land WHERE ST_Intersects(priogrid_land.centroid, gpw90.rast)) >>AS gpw90 >>WHERE gpw90.gid = 139303 >>GROUP BY gid,((gpw90.gpw90val)) >>; >> >>Gives: >>gid; gpw90 >>139303;39849 >> >>The value in the original raster is: >>39849.2 >> >>Question is then. How can i ensure that raster remain decimal and not integer >>after this import and >>query process? >> >>Thanks. >>Andreas > _______________________________________________ > postgis-users mailing list > [email protected]<mailto:[email protected]> > http://postgis.refractions.net/mailman/listinfo/postgis-users > > > _______________________________________________ > postgis-users mailing list > [email protected]<mailto:[email protected]> > http://postgis.refractions.net/mailman/listinfo/postgis-users > > > > _______________________________________________ > postgis-users mailing list > [email protected]<mailto:[email protected]> > http://postgis.refractions.net/mailman/listinfo/postgis-users > > > > _______________________________________________ > postgis-users mailing list > [email protected] > http://postgis.refractions.net/mailman/listinfo/postgis-users > _______________________________________________ postgis-users mailing list [email protected] http://postgis.refractions.net/mailman/listinfo/postgis-users
