(fix): pandas extension array repr for int64[pyarrow]

[ilan-gold]

Similar to #10315, we uncover more unexpected behavior with int64[pyarrow] as the testing data type.

• We were apparently relying on a call to np.array (via the xp.reshape) internally for the actual printing of the elements but this is not implemented for int64[pyarrow] (unlike other extension array types) and therefore broke without an explicit implementation. So we add that implementation

• Once this has been handled, the PandasExtensionArray is no longer being converted into a numpy array silently, and thus we need to ensure that the repr can handle the PandasExtensionArray natively.

• Lastly we add a KeyError to the lookup to ensure that this sort of thing cannot happen again.

• From (fix): no fill_value on reindex #10304

• Tests added

• User visible changes (including notable bug fixes) are documented in whats-new.rst

• New functions/methods are listed in api.rst

[ilan-gold]

Separately, I am not sure if max_width was meant to be inclusive or exclusive. The line length for the IntervalArray is exactly 100.

[ilan-gold]

Reverted the addition of IntervalArraya23de74 - see failure at https://github.com/pydata/xarray/actions/runs/15000671362/job/42146523169 for test_repr where the line length is exactly 100 instead of less than 100. I'm not clear what the behavior should be but let's keep this as minimal as possible. I will follow up with a more general "more test cases" PR.

[ilan-gold]

Is it possible these three tests are flaky? It seems to pass locally for me and on the face of it, has nothing to do with this PR>

[kmuehlbauer]

Is it possible these three tests are flaky? It seems to pass locally for me and on the face of it, has nothing to do with this PR>

@ilan-gold This surfaced with the release of dask 2025.5.0. Something has changed on the dask side and/or on the xarray side, which let the error surface.

Here is the stacktrace on my machine:

xarray/tests/test_backends.py:827 (TestDask.test_outer_indexing_reversed)
self=<xarray.tests.test_backends.TestDaskobjectat0x7f2da44c2ff0>deftest_outer_indexing_reversed(self) ->None:
# regression test for GH6560ds=xr.Dataset(
{"z": (("t", "p", "y", "x"), np.ones((1, 1, 31, 40)))},
)
withself.roundtrip(ds) ason_disk:
>subset=on_disk.isel(t=[0], p=0).z[:, ::10, ::10][:, ::-1, :]
/home/kai/python/projects/xarray/xarray/tests/test_backends.py:835: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ /home/kai/python/projects/xarray/xarray/core/dataset.py:2764: iniselvar=var.isel(var_indexers)
/home/kai/python/projects/xarray/xarray/core/variable.py:1032: iniselreturnself[key]
/home/kai/python/projects/xarray/xarray/core/variable.py:782: in__getitem__returnself._finalize_indexing_result(dims, data)
/home/kai/python/projects/xarray/xarray/core/variable.py:786: in_finalize_indexing_resultreturnself._replace(dims=dims, data=data)
/home/kai/python/projects/xarray/xarray/core/variable.py:938: in_replacereturntype(self)(dims, data, attrs, encoding, fastpath=True)
/home/kai/python/projects/xarray/xarray/core/variable.py:365: in__init__super().__init__(
/home/kai/python/projects/xarray/xarray/namedarray/core.py:264: in__init__self._dims=self._parse_dimensions(dims)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ self=<[AttributeError("'Variable' object has no attribute '_dims'") raisedinrepr()] Variableobjectat0x7f2d5c6c6380>dims= ('t', 'y', 'x')
def_parse_dimensions(self, dims: _DimsLike) ->_Dims:
dims= (dims,) ifisinstance(dims, str) elsetuple(dims)
iflen(dims) !=self.ndim:
>raiseValueError(
f"dimensions {dims} must have the same length as the "f"number of data dimensions, ndim={self.ndim}"
)
EValueError: dimensions ('t', 'y', 'x') musthavethesamelengthasthenumberofdatadimensions, ndim=4/home/kai/python/projects/xarray/xarray/namedarray/core.py:508: ValueError

[ilan-gold]

@kmuehlbauer Thanks for the heads up, sounds like I need to update my env then. Thanks for the info here!

[kmuehlbauer]

@kmuehlbauer Thanks for the heads up, sounds like I need to update my env then. Thanks for the info here!

See new issue #10321.