Hi,
I want to tile LiDAR data. My project contains about 550 tiles (1 km × 1 km), so processing everything at once in a standard FME workspace is not practical, as my PC would struggle with the memory requirements.
At the moment, I use a shapefile that contains a "to_run" attribute. In QGIS, I select a group of tiles and update this attribute. In FME, I use a Tester together with a Feature Reader so that only the selected tiles are processed.
While this works, the workspace still consumes a lot of memory, and I have to monitor the process more closely than I would like.
I think a solution using WorkspaceRunner could be more efficient, as it would allow FME to process one tile at a time instead of loading a large batch at once.
I have:
- A shapefile containing all tiles that need to be generated.
- The LiDAR source data.
What I am struggling with is that I don’t know how to do it exactly.
I feed the workspacerunner the shapefile with tiles but then in the bench that needs to run I also load in the shapefile to do the actual clipping? How does the runner know which tile it needs to use to clip?
Is something like this ok for the workspace to run? Feature reader has just the lidar dataset location and the spatial filter on intersect. Clipper only merge attributes and the combiner has a group by on “name” (an attribute from the shapefile).
In the workspacerunner I also need to give the source las data… Is it the same as in the featurereader or should I not work with a featurereader?
Any tips or suggestions would be greatly appreciated.
Thanks!