Hi

We fly over cities with a lidar. So we have multiple las files of a city. Because they don't start and stop on the city boundary we need to clip these files.

It's a super simple workbench, but it runs for 8 hours on average. Depends on the city offcourse.

I think it has a 50% chance of succes.

We get 2 types of errors.

The first one that says there isn't enough disk space or we don't have the right privileges.

The other one is that one of the files could not be written because the data is corrupt.

Altough if we don't change anything and run it again it completes with succes.

So we don't know what the actual problem is, maybe the network...

So If we don't know how to solve that, we can at least try to get a better chance of succes.

Would it be quicker to first select the files that actually intersect with the city boundary and only send those through the clipper? (Files that are completely withing the boundary doesn't need clipping).

At the moment I use the bouding box replacer for the las files in the spatialfilter.

The workbench looks like this:

Better would be to use the actual boundaries of the las files and not he bounding box. At the moment I don't know how to do this in FME.

I can do it in with python (pdal) so I have an sqlite that contains the actual boundaries.

(pdal tindex create "path\\cityname.sqlite" path\\*.las" -f "SQLite" --lyr_name "cityname")

So 2 big questions:

Is sending a file through the clipper without it actually being clipped and then writing it as fast as just writing the file?

How can we make the actual boundaries of a las file instead of the boundingbox?