Patrick Hoefler

Results 345 comments of Patrick Hoefler

Naming the Index Column

I think we should add ``rename_axis`` to dask/dask as well (I already added it to dask-expr), it's convenient in method chains

Naming the Index Column

We added rename_axis and df.index = df.index.rename("foo") also works with dask-expr, making df.index.name = "something" work is not really doable since we don't have mutable objects, so closing here

Incompatibility with package db-dtypes results in casts of object to string types

Yep we started converting object to string by default in 2023.07.1, you can turn this off through ``` dask.config.set({"dataframe.convert-string": False}) ```

Document that describe is using approximate algorithms

Feel free to work on this, there is no PR yet

BUG: unstack with sort=False fails when used with the level parameter…

One comment about the test, the tests should stay, just removing the assertion on shares memory is the way to go

BUG: df.groupby(col, as_index=False).value_counts() returns a DataFrame but is annotated as Series

Yep, this happens probably for every function where a Series is possible

BUG: df.groupby(col, as_index=False).value_counts() returns a DataFrame but is annotated as Series

If doable, I would obviously prefer 2 :)

Regression: Series.apply is too strict, does not allow to return a DateOffset

Thanks!

Release

there are wheels on Pypi, i.e. https://pypi.org/project/crick/#history Would be great if you could look into the pytest failures, that will certainly help

Release

@jrbourbeau could you look into he GitHub release thing?