Address pandas-related upstream errors in #10973

xarray/tests/test_cftimeindex.py

@@ -27,6 +27,7 @@
     assert_array_equal,
     assert_identical,
     has_cftime,
+    has_pandas_3,
     requires_cftime,
 )
 
@@ -960,8 +961,17 @@ def test_cftime_datetime_sub_cftimeindex(calendar):
 @pytest.mark.parametrize("calendar", _CFTIME_CALENDARS)
 def test_distant_cftime_datetime_sub_cftimeindex(calendar):
     a = xr.date_range("2000", periods=5, calendar=calendar, use_cftime=True)
-    with pytest.raises(ValueError, match="difference exceeds"):
-        a.date_type(1, 1, 1) - a
+    if not has_pandas_3:
+        with pytest.raises(ValueError, match="difference exceeds"):
+            a.date_type(1, 1, 1) - a
+    else:
+        result = a.date_type(1, 1, 1) - a
+        assert isinstance(result, pd.TimedeltaIndex)
+        assert result.unit == "us"
+
+        # Check that we can recover original index from subtracting timedeltas
+        roundtrip = CFTimeIndex(a.date_type(1, 1, 1) - result.to_pytimedelta())
+        assert roundtrip.equals(a)
 
 
 @requires_cftime

xarray/tests/test_coding_times.py

@@ -599,7 +599,7 @@ def test_cf_timedelta(timedeltas, units, numbers) -> None:
     if timedeltas == "NaT":
         timedeltas = np.timedelta64("NaT", "ns")
     else:
-        timedeltas = pd.to_timedelta(timedeltas).to_numpy()
+        timedeltas = pd.to_timedelta(timedeltas).as_unit("ns").to_numpy()
     numbers = np.array(numbers)
 
     expected = numbers
@@ -623,8 +623,9 @@ def test_cf_timedelta_2d() -> None:
     units = "days"
     numbers = np.atleast_2d([1, 2, 3])
 
-    timedeltas = np.atleast_2d(pd.to_timedelta(["1D", "2D", "3D"]).to_numpy())
-    expected = timedeltas
+    timedeltas = pd.to_timedelta(["1D", "2D", "3D"]).as_unit("ns")
+    timedeltas_2d = np.atleast_2d(timedeltas.to_numpy())
+    expected = timedeltas_2d
 
     actual = decode_cf_timedelta(numbers, units)
     assert_array_equal(expected, actual)

xarray/tests/test_groupby.py

@@ -3815,7 +3815,7 @@ def test_mean_with_cftime_objects_dask():
 
 def test_groupby_bins_datetime_mean():
     """Test groupby_bins with datetime mean (issue #6995)"""
-    times = pd.date_range("2020-01-01", "2020-02-01", freq="1h")
+    times = pd.date_range("2020-01-01", "2020-02-01", freq="1h", unit="ns")
     index = np.arange(len(times))
     bins = np.arange(0, len(index), 5)
 
@@ -3836,7 +3836,10 @@ def test_groupby_bins_mean_time_series():
     ds = xr.Dataset(
         {
             "measurement": ("trial", np.arange(0, 100, 10)),
-            "time": ("trial", pd.date_range("20240101T1500", "20240101T1501", 10)),
+            "time": (
+                "trial",
+                pd.date_range("20240101T1500", "20240101T1501", 10, unit="ns"),
+            ),
         }
     )
 

xarray/tests/test_interp.py

@@ -721,7 +721,7 @@ def test_datetime(x_new, expected) -> None:
     da = xr.DataArray(
         np.arange(24),
         dims="time",
-        coords={"time": pd.date_range("2000-01-01", periods=24)},
+        coords={"time": pd.date_range("2000-01-01", periods=24, unit="ns")},
     )
 
     actual = da.interp(time=x_new)
@@ -739,7 +739,7 @@ def test_datetime_single_string() -> None:
     da = xr.DataArray(
         np.arange(24),
         dims="time",
-        coords={"time": pd.date_range("2000-01-01", periods=24)},
+        coords={"time": pd.date_range("2000-01-01", periods=24, unit="ns")},
     )
     actual = da.interp(time="2000-01-01T12:00")
     expected = xr.DataArray(0.5)

xarray/tests/test_plot.py

@@ -2897,7 +2897,7 @@ def test_datetime_hue(self) -> None:
         ds2["hue"] = pd.date_range("2000-1-1", periods=4)
         ds2.plot.scatter(x="A", y="B", hue="hue")
 
-        ds2["hue"] = pd.timedelta_range("-1D", periods=4, freq="D")
+        ds2["hue"] = pd.timedelta_range("-1D", periods=4, freq="D", unit="ns")  # type: ignore[call-arg,unused-ignore]
         ds2.plot.scatter(x="A", y="B", hue="hue")
 
     def test_facetgrid_hue_style(self) -> None:

xarray/tests/test_variable.py

@@ -74,11 +74,16 @@ def var():
     ],
 )
 def test_as_compatible_data_writeable(data):
-    pd.set_option("mode.copy_on_write", True)
+    # In pandas 3 the mode.copy_on_write option defaults to True, so the option
+    # setting logic can be removed once our minimum version of pandas is
+    # greater than or equal to 3.
+    if not has_pandas_3:
+        pd.set_option("mode.copy_on_write", True)
     # GH8843, ensure writeable arrays for data_vars even with
     # pandas copy-on-write mode
     assert as_compatible_data(data).flags.writeable
-    pd.reset_option("mode.copy_on_write")
+    if not has_pandas_3:
+        pd.reset_option("mode.copy_on_write")
 
 
 class VariableSubclassobjects(NamedArraySubclassobjects, ABC):
@@ -225,7 +230,7 @@ def test_index_0d_timedelta64(self):
             x, np.timedelta64(td), np.dtype("timedelta64[us]")
         )
 
-        x = self.cls(["x"], pd.to_timedelta([td]))
+        x = self.cls(["x"], pd.to_timedelta([td]).as_unit("ns"))
         self._assertIndexedLikeNDArray(x, np.timedelta64(td), "timedelta64[ns]")
 
     def test_index_0d_not_a_time(self):
@@ -279,7 +284,7 @@ def test_0d_time_data(self):
         expected = np.datetime64("2000-01-01", "ns")
         assert x[0].values == expected
 
-    dt64_data = pd.date_range("1970-01-01", periods=3)
+    dt64_data = pd.date_range("1970-01-01", periods=3, unit="ns")
 
     @pytest.mark.parametrize(
         "values, unit",
@@ -311,7 +316,7 @@ def test_datetime64_conversion(self, values, unit):
             (td64_data.values.astype("timedelta64[m]"), "s"),
             (td64_data.values.astype("timedelta64[s]"), "s"),
             (td64_data.values.astype("timedelta64[ps]"), "ns"),
-            (td64_data.to_pytimedelta(), "ns"),
+            (td64_data.to_pytimedelta(), "us" if has_pandas_3 else "ns"),
         ],
     )
     def test_timedelta64_conversion(self, values, unit):
@@ -1109,8 +1114,8 @@ def test_datetime64_conversion_scalar(self, values, unit):
             (np.timedelta64(1, "m"), "s"),
             (np.timedelta64(1, "D"), "s"),
             (np.timedelta64(1001, "ps"), "ns"),
-            (pd.Timedelta("1 day"), "ns"),
-            (timedelta(days=1), "ns"),
+            (pd.Timedelta("1 day").as_unit("ns"), "ns"),
+            (timedelta(days=1), "us" if has_pandas_3 else "ns"),
         ],
     )
     def test_timedelta64_conversion_scalar(self, values, unit):
@@ -1135,7 +1140,8 @@ def test_0d_datetime(self):
         assert v.values == np.datetime64("2000-01-01", expected_unit)  # type: ignore[call-overload]
 
     @pytest.mark.parametrize(
-        "values, unit", [(pd.to_timedelta("1s"), "ns"), (np.timedelta64(1, "s"), "s")]
+        "values, unit",
+        [(pd.to_timedelta("1s").as_unit("ns"), "ns"), (np.timedelta64(1, "s"), "s")],
     )
     def test_0d_timedelta(self, values, unit):
         # todo: check, if this test is OK
@@ -3158,7 +3164,7 @@ def test_from_pint_wrapping_dask(self, Var):
         (np.datetime64("2000-01-01", "s"), "s"),
         (np.array([np.datetime64("2000-01-01", "ns")]), "ns"),
         (np.array([np.datetime64("2000-01-01", "s")]), "s"),
-        (pd.date_range("2000", periods=1), "ns"),
+        (pd.date_range("2000", periods=1, unit="ns"), "ns"),
         (
             datetime(2000, 1, 1),
             "us" if has_pandas_3 else "ns",
@@ -3167,10 +3173,17 @@ def test_from_pint_wrapping_dask(self, Var):
             np.array([datetime(2000, 1, 1)]),
             "us" if has_pandas_3 else "ns",
         ),
-        (pd.date_range("2000", periods=1, tz=pytz.timezone("America/New_York")), "ns"),
+        (
+            pd.date_range(
+                "2000", periods=1, tz=pytz.timezone("America/New_York"), unit="ns"
+            ),
+            "ns",
+        ),
         (
             pd.Series(
-                pd.date_range("2000", periods=1, tz=pytz.timezone("America/New_York"))
+                pd.date_range(
+                    "2000", periods=1, tz=pytz.timezone("America/New_York"), unit="ns"
+                )
             ),
             "ns",
         ),
@@ -3245,8 +3258,8 @@ def test_pandas_two_only_datetime_conversion_warnings(
         (np.array([np.timedelta64(10, "ns")]), "ns"),
         (np.array([np.timedelta64(10, "s")]), "s"),
         (pd.timedelta_range("1", periods=1), "ns"),
-        (timedelta(days=1), "ns"),
-        (np.array([timedelta(days=1)]), "ns"),
+        (timedelta(days=1), "us" if has_pandas_3 else "ns"),
+        (np.array([timedelta(days=1)]), "us" if has_pandas_3 else "ns"),
         (pd.timedelta_range("1", periods=1).astype("timedelta64[s]"), "s"),
     ],
     ids=lambda x: f"{x}",