Skip to content

GH-35650: [Python] Cast array fill values in fill_null instead of as_py() #50890

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

Open
Nishuuzz wants to merge 1 commit into
base: main
Choose a base branch
from
+30 −1
Open

GH-35650: [Python] Cast array fill values in fill_null instead of as_py() #50890

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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: 3 additions & 1 deletion python/pyarrow/compute.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -562,7 +562,9 @@ def fill_null(values, fill_value):
if not isinstance(fill_value, (pa.Array, pa.ChunkedArray, pa.Scalar)):
fill_value = pa.scalar(fill_value, type=values.type)
elif values.type != fill_value.type:
fill_value = pa.scalar(fill_value.as_py(), type=values.type)
# Cast rather than going through as_py(), which only exists on Scalar
# and so failed with an AttributeError for array fill values.
fill_value = fill_value.cast(values.type)

return call_function("coalesce", [values, fill_value])

Expand Down
27 changes: 27 additions & 0 deletions python/pyarrow/tests/test_compute.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2112,6 +2112,33 @@ def test_fill_null():
assert result.equals(expected)


def test_fill_null_array_different_type():
# GH-35650: an array or chunked array fill value whose type differs from
# the values is cast, rather than raising AttributeError because only
# Scalar has an as_py() method.
arr = pa.array([1, 2, None, 4, None], type=pa.int64())

fill_value = pa.array([10, 20, 30, 40, 50], type=pa.int32())
result = arr.fill_null(fill_value)
expected = pa.array([1, 2, 30, 4, 50], type=pa.int64())
assert result.equals(expected)

chunked_fill = pa.chunked_array([[10, 20], [30, 40, 50]], type=pa.int32())
result = pa.chunked_array([arr]).fill_null(chunked_fill)
assert result.equals(pa.chunked_array([expected]))

# the reported case: a fixed size binary fill value for a string array
values = pa.array(['ab', None], type=pa.string())
fill_value = pa.array([b'cd', b'ef'], type=pa.binary(2))
result = values.fill_null(fill_value)
assert result.equals(pa.array(['ab', 'ef'], type=pa.string()))

# the result keeps the type of `values`, so a fill value that cannot be
# cast to it is an error rather than promoting the result
with pytest.raises(pa.ArrowInvalid):
arr.fill_null(pa.array([1.5, 2.5, 3.5, 4.5, 5.5], type=pa.float64()))


@pytest.mark.parametrize('arrow_type', numerical_arrow_types)
def test_fill_null_array(arrow_type):
arr = pa.array([1, 2, None, 4], type=arrow_type)
Expand Down