Skip to main content
12 events
when toggle format what by license comment
Aug 6, 2020 at 8:38 comment added Theo F @dr_jts so after a week or so of the script running, it has finished. Most of the country I used 5km grid squares to iterate through but had to use 1km squares for urban areas where the process was struggling. In some urban areas a 5km grid would take 1 hour, but the equivalent 25x1km grids would take 8mins.
Jul 24, 2020 at 19:46 comment added Theo F @dr_jts the pipeline is up and running. I'm using square 10x10km grids and iterating through them in a loop. I might have to reduce the grids to 5x5km as even at 10km the process can slow and even close server connection (Argh!) on more complex areas of land with many polygons.
Jul 24, 2020 at 19:44 vote accept Theo F
Jul 23, 2020 at 12:01 history edited Theo F CC BY-SA 4.0
added 163 characters in body
Jul 23, 2020 at 10:47 history edited Theo F CC BY-SA 4.0
added 482 characters in body
Jul 23, 2020 at 10:12 history edited Theo F CC BY-SA 4.0
added new solution
Jul 23, 2020 at 0:10 comment added dr_jts When looking to scale spatial processing queries up and/or out, gridding is usually the first technique to explore.
Jul 23, 2020 at 0:10 comment added dr_jts Good stuff, would be interesting to hear how that works.
Jul 22, 2020 at 18:30 comment added Theo F @dr_jts thanks, and this is actually what I've planned to do. Using 5kmx5km grid squares for the entire country and iterating through the grids by a grid id. This will be done in python (using psycopg2).
Jul 22, 2020 at 18:19 comment added dr_jts To scale this up/out, you could process batches of polygons using a rectangular grid defined over the data space. The constructed gap polygons can be clipped to grid cells. and optional unioned afterwards.
Jul 22, 2020 at 15:21 history edited Theo F CC BY-SA 4.0
added 51 characters in body
Jul 22, 2020 at 14:45 history answered Theo F CC BY-SA 4.0