fix: pd.Series in pandas>=3 does not preserve object dtype metadata

xarray/core/variable.py

@@ -13,6 +13,7 @@
 import numpy as np
 import pandas as pd
 from numpy.typing import ArrayLike
+from packaging.version import Version
 
 import xarray as xr  # only for Dataset and DataArray
 from xarray.compat.array_api_compat import to_like_array
@@ -208,7 +209,10 @@ def _maybe_wrap_data(data):
 
 def _possibly_convert_objects(values):
     """Convert object arrays into datetime64 and timedelta64 according
-    to the pandas convention.
+    to the pandas convention.  For backwards compat, as of 3.0.0 pandas,
+    object dtype inputs are cast to strings by `pandas.Series`
+    but we output them as object dtype with the input metadata preserved as well.
+
 
     * datetime.datetime
     * datetime.timedelta
@@ -223,6 +227,18 @@ def _possibly_convert_objects(values):
             result.flags.writeable = True
         except ValueError:
             result = result.copy()
+    # For why we need this behavior: https://github.com/pandas-dev/pandas/issues/61938
+    # Object datatype inputs that are strings
+    # will be converted to strings by `pandas.Series`, and as of 3.0.0, lose
+    # metadata.  If the roundtrip back to numpy in this function yields an
+    # object array again, the dtype.metadata will be preserved.
+    if (
+        result.dtype.kind == "O"
+        and values.dtype.kind == "O"
+        and Version(pd.__version__) >= Version("3.0.0dev0")
+    ):
+        # need to copy to be able to override `dtype`
+        result.dtype = values.dtype
     return result
 
 

xarray/tests/test_dataset.py

@@ -13,6 +13,7 @@
 import numpy as np
 import pandas as pd
 import pytest
+from packaging.version import Version
 from pandas.core.indexes.datetimes import DatetimeIndex
 
 # remove once numpy 2.0 is the oldest supported version
@@ -299,7 +300,7 @@ def test_repr(self) -> None:
                 var1     (dim1, dim2) float64 576B -0.9891 -0.3678 1.288 ... -0.2116 0.364
                 var2     (dim1, dim2) float64 576B 0.953 1.52 1.704 ... 0.1347 -0.6423
                 var3     (dim3, dim1) float64 640B 0.4107 0.9941 0.1665 ... 0.716 1.555
-                var4     (dim1) category 32B b c b a c a c a{var5}
+                var4     (dim1) category 3{6 if Version(pd.__version__) >= Version("3.0.0dev0") else 2}B b c b a c a c a{var5}
             Attributes:
                 foo:      bar"""
         )