Pieter Roggemans

> I'm pretty sure it's not environment related. Have you tried it locally? Did I misunderstand that if you used the steps I posted it was working properly?

> I was basically asking whether you had tried replicating the steps for item 2. above. I gave it a quick try, but wasn't able to replicate. I wouldn't expect...

I've never used geoparquet... but possibly using `read_file` will support this case already... The gdal documentation lists some different geometry column types that would be supported: https://gdal.org/en/latest/drivers/vector/parquet.html

I'm not able to reproduce, but I'm using windows, so that's a significant difference. Is there a specific reason you are installing pyogrio via pip instead of via conda?

You can filter it like any warning, e.g. like this: ``` python import warnings import pyogrio file = "C:\Temp\polygon_parcel\polygon-parcel.gpkg" with open(file, "rb") as f: with warnings.catch_warnings(): warnings.filterwarnings("ignore", ".*has GPKG application_id.*",...

> While I understand, I think we're way past the time to have this discussion. This decision predates my involvement so don't know the reasons but my guess would ne...

> To be fair, I am not that convinced that our current behaviour is the worse one of the two. I understand that the model you suggest is what GIS...

@m-richards Now we are with three agreeing (that the impact of changing this would be a too large break of API). Hopefully someone else can think of a solution still...

> I don't think we should try to get closer to traditional GIS I don't have significant practical experience in using "traditional" GIS programming interfaces (I suppose you mean things...

> I thought about this for a while and could not come with a use case when I would want to return the whole GeoDataFrame. I just don't see when...