Even Rouault wrote: > Le vendredi 20 juillet 2012 20:58:08, Sven Geggus a écrit : >> Even Rouault <[email protected]> wrote: >> > Testing with larger areas, like whole France or Europe, shows sluggish >> > performance when ways are built from nodes, but that's perhaps >> expected. >> > I didn't compare with other tools to know if the indexing or request >> > strategy is particularly bad. >> >> Hm, would it be possible with this to e.g. convert all the forests >> from a complete planetfile to a polygon shapefile while discarding >> all the rest? > > Yes, this should be possible by specifying the appropriate -where clause > to > ogr2ogr. However this will have roughly the same performance as completely > converting the whole planetfile, so it will take some time... > > Hum, actually while writing this, I see that the attribute filter is > evaluated > after having reconstructed the geometry. So there's perhaps something I > can do > to speed that up a bit, in order to evaluate the attribute filter before > and > skip the geometry building if the attribute filter rejects the filter.
I made a test with germany.osm.pbf file from Geofabrik. I used GDAL development version r24717 and my ogr2ogr command was: ogr2ogr -f SQlite -dsco spatialite=yes germany_forests.sqlite germany.osm.pbf polygons multipolygons -gt 20000 --config OGR_SQLITE_SYNCHRONOUS OFF -where "landuse='forest'" -progress Execution took 12 minutes. This much landuse=forest areas were found: Layer name: polygons Geometry: Polygon Feature Count: 262683 Layer name: multipolygons Geometry: Multi Polygon Feature Count: 6917 There are some more timings and info about the hardware used in the test in this mail http://article.gmane.org/gmane.comp.gis.gdal.devel/32175 -Jukka Rahkonen- _______________________________________________ dev mailing list [email protected] http://lists.openstreetmap.org/listinfo/dev
