Convert to compatible NumPy dtype for MaskedArray to_numpy #55058
Conversation
mroeschke
added
the
NA - MaskedArrays
|
I think this is pretty reasonable. Is this part of a larger set of changes or is this self-contained? Probably worth a whatsnew |
|
This is something we wanted to clean up for a while now, but yes I have to add a whatsnew (didn't get to it yet because this needs more than a single line) |
|
Added the whatsnew |
|
Does this close #48891? |
|
yep |
# Conflicts: # pandas/tests/base/test_conversion.py # pandas/tests/frame/test_repr.py
|
cc @mroeschke I think this is ready for a review |
pandas/io/formats/format.py
Outdated
| @@ -1527,6 +1528,8 @@ def _format_strings(self) -> list[str]: | |||
| if isinstance(values, Categorical): | |||
| # Categorical is special for now, so that we can preserve tzinfo | |||
| array = values._internal_get_values() | |||
| elif isinstance(values, BaseMaskedArray): | |||
| array = values.to_numpy(na_value=NA) | |||
There was a problem hiding this comment.
As I discovered in https://github.com/pandas-dev/pandas/pull/56003/files#diff-c2df0eefbdda16cb85d5e0a305db90101f92bfb790a8e989ab2e99b935b7f6b3, maybe we just want to do to_numpy(dtype=object) here?
There was a problem hiding this comment.
Yep hat should work as well
| @pytest.mark.parametrize("na_action", [None, "ignore"]) | ||
| def test_map(self, data_missing, na_action): | ||
| result = data_missing.map(lambda x: x, na_action=na_action) | ||
| if data_missing.dtype == Float32Dtype(): |
There was a problem hiding this comment.
If there a solution to avoid this?
There was a problem hiding this comment.
Not with roundtripping through object, maybe your other suggestion above solves this.
No if your suggestion doesn't work
# Conflicts: # doc/source/whatsnew/v2.2.0.rst
|
Can we get this in if ci is green now? Just merged main to resolve whatsnew conflicts |
Co-authored-by: Matthew Roeschke <10647082+mroeschke@users.noreply.github.com>
|
Thanks @phofl |
doc/source/whatsnew/vX.X.X.rstfile if fixing a bug or adding a new feature.This still needs a whatsnew