Skip to content

BUG: .quantile with empty temporal data #63407

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
mroeschke merged 5 commits into from
Dec 18, 2025
+24 −2
Merged

BUG: .quantile with empty temporal data #63407

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
5486e75
BUG: .quantile with empty temporal data
mroeschke Dec 18, 2025
85fcf10
Merge remote-tracking branch 'upstream/main' into bug/quantile/temporal
mroeschke Dec 18, 2025
a56a87d
Revert old change, ensure accounts for .unit
mroeschke Dec 18, 2025
f7238b3
Merge remote-tracking branch 'upstream/main' into bug/quantile/temporal
mroeschke Dec 18, 2025
b7570be
Add DataFrame version of the test
mroeschke Dec 18, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 1 addition & 3 deletions pandas/core/arrays/_mixins.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -512,9 +512,7 @@ def _quantile(
else:
# e.g. test_quantile_empty we are empty integer dtype and res_values
# has floating dtype
# TODO: technically __init__ isn't defined here.
# Should we raise NotImplementedError and handle this on NumpyEA?
return type(self)(res_values) # type: ignore[call-arg]
return type(self)._from_sequence(res_values) # type: ignore[call-arg]
Copy link
Member

@jorisvandenbossche jorisvandenbossche Dec 18, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

While having an else clause that actually cannot work doesn't seem great (and so this change might be good anyway), looking at the code above (if res_values.dtype == self._ndarray.dtype:), I am wondering if the problem might not be that quantile_with_mask is not returning the correct dtype in case of empty data.

In that case it returns NaT in an array, and it seems to return nanoseconds, which seems wrong?

Copy link
Member

@jorisvandenbossche jorisvandenbossche Dec 18, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looking a bit more, quantile_with_mask is returning nanoseconds in this case because of the fill_value being passed, which is self._internal_fill_value, and that seems to be defined as np.timedelta64('NaT','ns'). That probably need to take the unit of self into account?

Copy link
Member Author

@mroeschke mroeschke Dec 18, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Nice find. The problem was _internal_fill_value assuming nanoseconds for NaT.

I reverted the changes to ._quantile to just be conservative, though probably the right long term change


# ------------------------------------------------------------------------
# numpy-like methods
Expand Down
7 changes: 7 additions & 0 deletions pandas/tests/series/methods/test_quantile.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -245,3 +245,10 @@ def test_quantile_dtype_size(self, any_int_ea_dtype):
result = ser.quantile([0.1, 0.5])
expected = Series([1, 1], dtype=any_int_ea_dtype, index=[0.1, 0.5])
tm.assert_series_equal(result, expected)


@pytest.mark.parametrize("typ", ["datetime64", "timedelta64"])
def test_quantile_empty_datetimelike(typ, unit):
Copy link
Member

@jorisvandenbossche jorisvandenbossche Dec 18, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Maybe it would be good to also add a dataframe version of this test, such that there is an actual dtype to verify (to assert it has preserved the unit)

ser = Series([], dtype=f"{typ}[{unit}]")
result = ser.quantile()
assert result is pd.NaT
Loading