type align #1138
type align #1138
Conversation
pandas-stubs/core/frame.pyi
Outdated
| limit: int | None = ..., | ||
| fill_axis: Axis = ..., | ||
| broadcast_axis: Axis | None = ..., | ||
| fill_value: Hashable | None = ..., |
There was a problem hiding this comment.
some methods use Hashable for fill_value, others Scalar, I'm not sure. anyway for now in this PR I'm just going with what's in pandas
pandas-stubs/core/frame.pyi
Outdated
| limit: int | None = ..., | ||
| fill_axis: Axis = ..., | ||
| broadcast_axis: Axis | None = ..., | ||
| fill_value: Hashable | None = ..., |
There was a problem hiding this comment.
This should be Scalar | NAType | None . In the stubs, we want to promote "good" behavior, and Hashable is too wide. In addition, the docs say "fill value: scalar, default np.nan" . Maybe the pandas source should change??
pandas-stubs/core/series.pyi
Outdated
| limit: int | None = ..., | ||
| fill_axis: AxisIndex = ..., | ||
| broadcast_axis: AxisIndex | None = ..., | ||
| fill_value: Hashable | None = ..., |
|
By the way, hope it's OK to ask, is a release planned any time soon? Some of the recent improvements would help out with internal type checking in Narwhals (no hurry / pressure of course!) |
Yes. I think once your other PR's are done, I will do a release. I was going to do one this past Saturday, but had to wait for pyright to be updated, and by the time that happened, you had created a bunch of PRs!! |
@MarcoGorelli I just did a release today |
assert_type()to assert the type of any return value