[gdal-dev] ogr2ogr intersect very slow at command line with large sqlite input

CTL101 chrislloyd2 at yahoo.co.uk
Mon Apr 10 15:38:46 PDT 2017

Hi again Steve, The input is osm data, filtered so that only highways of some
significance remain. I am trying to get road intersections as points based
on certain criteria. I have an ogr2ogr sqlite based workflow that works on
test data. But scaling up is proving to be a problem for this particular
intersect operation, which occurs early on. Based on my source and sample
data i would say that there are possibly several million intersections
between the tables to compute. The intersection is between lines (roads),
but I am only interested in the point output where lines meet. If there is a
more efficient algorithm that i could use in ogr2ogr then i would be
grateful of advice.

I had not considered postgresql for this part of the process but would
prefer to streamline the workflow rather than create a new one if avoidable.
Might conversion of the sqlite tables to another format in ogr2ogr yield
significant performance dividends? Or do I need to resign myself to trying a
fresh approach?
Regards, Chris

View this message in context: http://osgeo-org.1560.x6.nabble.com/ogr2ogr-intersect-very-slow-at-command-line-with-large-sqlite-input-tp5316545p5316749.html
Sent from the GDAL - Dev mailing list archive at Nabble.com.

More information about the gdal-dev mailing list