Skip to content

Write chunks with negative zero values and a zero fill value #3216

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.

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
merged 8 commits into from
Aug 6, 2025
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
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
1 change: 1 addition & 0 deletions changes/3144.bugfix.rst
Original file line number Diff line number Diff line change
@@ -0,0 +1 @@
Ensure that -0.0 is not considered equal to 0.0 when checking if all the values in a chunk are equal to an array's fill value.```
9 changes: 9 additions & 0 deletions src/zarr/core/buffer/core.py
Original file line number Diff line number Diff line change
Expand Up @@ -523,6 +523,15 @@ def all_equal(self, other: Any, equal_nan: bool = True) -> bool:
if other is None:
# Handle None fill_value for Zarr V2
return False
# Handle positive and negative zero by comparing bit patterns:
if (
np.asarray(other).dtype.kind == "f"
and other == 0.0
and self._data.dtype.kind not in ("U", "S", "T", "O", "V")
):
_data, other = np.broadcast_arrays(self._data, np.asarray(other, self._data.dtype))
void_dtype = "V" + str(_data.dtype.itemsize)
return np.array_equal(_data.view(void_dtype), other.view(void_dtype))
# use array_equal to obtain equal_nan=True functionality
# Since fill-value is a scalar, isn't there a faster path than allocating a new array for fill value
# every single time we have to write data?
Expand Down
24 changes: 24 additions & 0 deletions tests/test_array.py
Original file line number Diff line number Diff line change
Expand Up @@ -870,6 +870,30 @@ def test_write_empty_chunks_behavior(
assert arr.nchunks_initialized == arr.nchunks


@pytest.mark.parametrize("store", ["memory"], indirect=True)
@pytest.mark.parametrize("fill_value", [0.0, -0.0])
@pytest.mark.parametrize("dtype", ["f4", "f2"])
def test_write_empty_chunks_negative_zero(
zarr_format: ZarrFormat, store: MemoryStore, fill_value: float, dtype: str
) -> None:
# regression test for https://github.com/zarr-developers/zarr-python/issues/3144

arr = zarr.create_array(
store=store,
shape=(2,),
zarr_format=zarr_format,
dtype=dtype,
fill_value=fill_value,
chunks=(1,),
config={"write_empty_chunks": False},
)
assert arr.nchunks_initialized == 0

# initialize the with the negated fill value (-0.0 for +0.0, +0.0 for -0.0)
arr[:] = -fill_value
assert arr.nchunks_initialized == arr.nchunks
Copy link
Contributor

Choose a reason for hiding this comment

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

this test is fine but ideally we would be testing the altered function explicitly, instead of indirectly via array creation + chunk writing. this is not a blocker for this PR, just something to sort out down the road

Copy link
Contributor Author

Choose a reason for hiding this comment

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

That test is basically copied from test_write_empty_chunks_behavior right above it. But yeah, it might be worth to have both a unit and an integration test in this case (:



@pytest.mark.parametrize(
("fill_value", "expected"),
[
Expand Down
Loading