Skip to content

BUG: Fix drop(index=[pd.NA]) for PyArrow-backed Index (GH#63304) #63910

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

Closed
dedlYTiTAN wants to merge 8 commits into from
+65 −1
Closed

BUG: Fix drop(index=[pd.NA]) for PyArrow-backed Index (GH#63304) #63910

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
03a5ec2
FIX: drop(index=[pd.NA]) fails for Arrow-backed index (#63304)
Jan 28, 2026
4c7b324
TST: Skip Arrow drop test if pyarrow < 13.0.0
Jan 28, 2026
eb0433d
STY: linting fixes (isort, whitespace) for #63304 test
Jan 28, 2026
3673427
STY: Apply ruff format to test_gh63304.py
Jan 28, 2026
453f450
FIX: restrict null-casting in isin to explicit pd.NA (fix regression)
Jan 28, 2026
a3d0d0e
STY: apply ruff fixes to arrow/array.py
Jan 28, 2026
00633a5
REF: replace lib.missing.NA with isna() to fix mypy error
Jan 28, 2026
4fd9f8c
Merge branch 'main' into fix-gh63304-arrow-drop-na-v2
dedlYTiTAN Jan 28, 2026
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
19 changes: 18 additions & 1 deletion pandas/core/arrays/arrow/array.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1418,7 +1418,24 @@ def isin(self, values: ArrayLike) -> npt.NDArray[np.bool_]:
if not len(values):
return np.zeros(len(self), dtype=bool)

result = pc.is_in(self._pa_array, value_set=pa.array(values))
value_set = pa.array(values, from_pandas=True)
if pa.types.is_null(value_set.type):
# GH#63304: If we have explicit pd.NA, we want to allow the comparison
# to return False (not found) rather than raising ArrowInvalid.
# However, we need to be careful not to swallow other types that might
# be inferred as null (e.g. [np.nan]) which logic elsewhere might rely
# on crashing to trigger fallback (e.g. in parsers).
has_pd_na = False
for x in values:
# GH#63304: Check for pd.NA (NAType) specifically
if isna(x) and not isinstance(x, (float, np.floating, type(None))):
has_pd_na = True
break

if has_pd_na:
value_set = value_set.cast(self._pa_array.type)

result = pc.is_in(self._pa_array, value_set=value_set)
# pyarrow 2.0.0 returned nulls, so we explicitly specify dtype to convert nulls
# to False
return np.array(result, dtype=np.bool_)
Expand Down
47 changes: 47 additions & 0 deletions pandas/tests/indexing/test_gh63304.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,47 @@
import pytest

import pandas as pd
import pandas._testing as tm

pa = pytest.importorskip("pyarrow", minversion="13.0.0")


def test_drop_na_arrow_index():
# GH#63304
# Test that dropping pd.NA from PyArrow-backed Index does not raise ArrowInvalid

# integer
df = pd.DataFrame(
{"A": [1, 2, 3]}, index=pd.Index([1, 2, 3], dtype="int64[pyarrow]")
)
# pd.NA is not in index, should raise KeyError, but NOT ArrowInvalid
with pytest.raises(KeyError, match="not found in axis"):
df.drop(index=[pd.NA])

# string
df = pd.DataFrame(
{"A": [1, 2, 3]}, index=pd.Index(["a", "b", "c"], dtype="string[pyarrow]")
)
with pytest.raises(KeyError, match="not found in axis"):
df.drop(index=[pd.NA])

# binary
df = pd.DataFrame(
{"A": [1, 2, 3]},
index=pd.Index([b"a", b"b", b"c"], dtype="binary[pyarrow]"),
)
with pytest.raises(KeyError, match="not found in axis"):
df.drop(index=[pd.NA])

# Case where NA IS in the index (should verify it drops correctly)
df = pd.DataFrame({"A": [1, 2]}, index=pd.Index([1, pd.NA], dtype="int64[pyarrow]"))
result = df.drop(index=[pd.NA])
expected = pd.DataFrame({"A": [1]}, index=pd.Index([1], dtype="int64[pyarrow]"))
tm.assert_frame_equal(result, expected)

df = pd.DataFrame(
{"A": [1, 2]}, index=pd.Index(["a", pd.NA], dtype="string[pyarrow]")
)
result = df.drop(index=[pd.NA])
expected = pd.DataFrame({"A": [1]}, index=pd.Index(["a"], dtype="string[pyarrow]"))
tm.assert_frame_equal(result, expected)
Loading