I'm a bit torn about this change - dask_geopandas can't actually load the files, so they'll be loaded into memory here regardless of this change. I just thought it might be easier to just stick to this lazy datatype, so that later functions don't always have to check whether something is a geopandas or a dask_geopandas dataframe.
Any thoughts on this?
Seems reasonable. I can't think of scenarios where dask_geopandas would not work at the moment, we will see in the future how we manage the new vector processes with it.
I'm a bit torn about this change - dask_geopandas can't actually load the files, so they'll be loaded into memory here regardless of this change. I just thought it might be easier to just stick to this lazy datatype, so that later functions don't always have to check whether something is a geopandas or a dask_geopandas dataframe. Any thoughts on this?