Split a pointcloud file over several RDD records instead of having it in one single record.

[romulogoncalves]

If we understood correctly, for each LAS/LAZ file loaded, geotrellis creates a RDD record. If a data set has some large files this will create unbalance task processing, i.e., ones will take a lot longer than others.

A solution would be to split large LAS/LAZ files into multiple RDD records. From PDAL we have the following filter functions: chipper, divider and splitter. We have tried them in a PDAL pipeline, but then when we check the number of elements in the RDD, they all report 1.

Is there away in GeoTrellis to split the LAZ file into multiple RDD records and run a PDAL pipeline?

We could do tilling but then how is possible to run a PDAL pipeline on each Tile?