Skip to content

Conversation

jorisvandenbossche
Copy link
Member

@jorisvandenbossche jorisvandenbossche commented Feb 12, 2022

I noticed that we were actually subclassing the legacy engine from dask, and not the newer pyarrow.dataset based one. This doesn't change anything for the rest in our code (all tests pass), and this should also get rid of a bunch of deprecation warnings (from pyarrow, from using the legacy APIs)

It might not yet be available for older dask versions, in which case I can add a fallback to subclass the legacy engine in that case.

@martinfleis
Copy link
Member

It might not yet be available for older dask versions, in which case I can add a fallback to subclass the legacy engine in that case.

Can we just bump requirements? We currently require a version from 6 Jun 2020.

@jorisvandenbossche
Copy link
Member Author

The current failure is actually not because of this PR, but so the Feather PR where I started using from dask.layers import DataFrameIOLayer.

Checking this new class, it seems to be introduced in dask/dask#7415, which corresponds to dask 2021.05.0

@jorisvandenbossche jorisvandenbossche merged commit 87d22f4 into geopandas:main Feb 15, 2022
@jorisvandenbossche jorisvandenbossche deleted the parquet-non-legacy branch February 15, 2022 08:10
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants