The Dask team is excited to announce the release of 2022.02.01 today! You can see the complete list of changes in the Changelog. Highlights include:
- Dask DataFrame has expanded Pandas coverage.
-
std()
now works for datetime Series -
pivot_table()
now supportsfirst
&last
aggregations.
Thanks to new contributors Ben Glossner and @nordange!
-
Dask clients now support preload scripts, making it easier to customize client initialization (thanks to @bryanweber). Previously this feature was only supported by the scheduler & workers.
-
When a Dask worker runs low on RAM, it will spill excess data to disk. With this release, you can now configure how much disk space a worker is allowed to use for storing excess data (thanks @ncclementi!). Configuration — Dask documentation
-
This release drops support for Python 3.7, in accordance with NEP29.
Thanks to @jsignell for managing this release, and to everyone who contributed across the Dask org. We had 27 contributors this release, including 6 first time contributors!