geopackage support/SQL queries in python

[SjoerdBraaksma]

Hi! Loving the package so far. I am currently writing all my old code using command-line pdal and capturing the output, into python-pdal functions.

I was wondering: Is it possible to, when loading a geopackage, use a synonym for the datasource name in the query? Especially when file location names become long, this can be tedious.

For example:

pipeline= pdal.Filter.overlay(
   dimension="dimension",
   datasource= "test/file/in/path/further/down/1234/gpkg",
   as= "testfile",
  query= "select classification from testfile where classification == 11",
  layer="roads"
).pipeline(data)

thanks in advance!

[hobu]

Is it possible to, when loading a geopackage, use a synonym for the datasource name in the query?

Can you show me an example of what you mean? I would think this would have to be supported by OGR, as PDAL's simply passing this down to OGR to select the geometry used for intersection.

[SjoerdBraaksma]

Hi!

yes, the above example shows what I mean. How I do it now:

pipeline= pdal.Filter.overlay(
   dimension="dimension",
   datasource= "test/file/in/path/further/down/1234/gpkg",
  query= "select classification from test/file/in/path/further/down/1234/gpkg where classification == 11",
  layer="roads"
).pipeline(data)

What I mean:

pipeline= pdal.Filter.overlay(
   dimension="dimension",
   datasource= "test/file/in/path/further/down/1234/gpkg",
   as= "testfile",
  query= "select classification from testfile where classification == 11",
  layer="roads"
).pipeline(data)

This would also possibly open the way to more complex SQL queries where you can combine multiple layers/datasets before passing it to overlay.

[hobu]

I would accept a PR to PDAL that implemented this substitution, but do know that the OGR_SQL variant supports table aliasing and a bunch of other stuff. This contribution would belong in the PDAL repository, so please reopen a PR there if you do implement it.

[SjoerdBraaksma]

I thought about it for a while, and I think the easiest and most clear solution is just by constructing an f-string:

pipeline= pdal.Filter.overlay(
   dimension="dimension",
   datasource= datasource,
   as= "testfile",
  query= f"select classification from {os.splitext(datasource)[0]} where classification == 11",
  layer="roads"
).pipeline(data)