Skip to content

Fix a regression with scalar indexing due to #1800 #1875

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 9 commits into from
May 16, 2024
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
2 changes: 2 additions & 0 deletions docs/release.rst
Original file line number Diff line number Diff line change
Expand Up @@ -31,6 +31,8 @@ Docs

Maintenance
~~~~~~~~~~~
* Fix a regression when getting or setting a single value from arrays with size-1 chunks.
By :user:`Deepak Cherian <dcherian>` :issue:`1874`

Deprecations
~~~~~~~~~~~~
Expand Down
4 changes: 3 additions & 1 deletion zarr/core.py
Original file line number Diff line number Diff line change
Expand Up @@ -2030,7 +2030,9 @@ def _process_chunk(
and not self._filters
and self._dtype != object
):
dest = out[out_selection]
# For 0D arrays out_selection = () and out[out_selection] is a scalar
# Avoid that
dest = out[out_selection] if out_selection else out
# Assume that array-like objects that doesn't have a
# `writeable` flag is writable.
dest_is_writable = getattr(dest, "writeable", True)
Expand Down
2 changes: 2 additions & 0 deletions zarr/indexing.py
Original file line number Diff line number Diff line change
Expand Up @@ -52,6 +52,8 @@ def is_scalar(value, dtype):
return True
if isinstance(value, tuple) and dtype.names and len(value) == len(dtype.names):
return True
if dtype.kind == "O" and not isinstance(value, np.ndarray):
Copy link
Contributor Author

@dcherian dcherian May 16, 2024

Choose a reason for hiding this comment

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

This is the core issue in #1874. We need to treat [1, 'C'] as a scalar for object arrays.

Is this the right condition?

cc @jakirkham @aplowman

return True
return False


Expand Down
49 changes: 49 additions & 0 deletions zarr/tests/test_core.py
Original file line number Diff line number Diff line change
Expand Up @@ -3157,3 +3157,52 @@ def test_issue_1279(tmpdir):

written_data = ds_reopened[:]
assert_array_equal(data, written_data)


def test_scalar_indexing():
store = zarr.KVStore({})

store["a"] = zarr.create((3,), chunks=(1,), store=store)
store["a"][:] = [1, 2, 3]

assert store["a"][1] == np.array(2.0)
assert store["a"][(1,)] == np.array(2.0)

store["a"][slice(1)] = [-1]
assert store["a"][0] == np.array(-1)

store["a"][0] = -2
assert store["a"][0] == np.array(-2)

store["a"][slice(1)] = (-3,)
assert store["a"][0] == np.array(-3)


def test_object_array_indexing():
# regression test for #1874
from numcodecs import MsgPack

root = zarr.group()
arr = root.create_dataset(
name="my_dataset",
shape=0,
dtype=object,
object_codec=MsgPack(),
)
new_items = [
["A", 1],
["B", 2, "hello"],
]
arr_add = np.empty(len(new_items), dtype=object)
arr_add[:] = new_items
arr.append(arr_add)

# heterogeneous elements
elem = ["C", 3]
arr[0] = elem
assert arr[0] == elem

# homogeneous elements
elem = [1, 3]
arr[1] = elem
assert arr[1] == elem