adapt more tests to the copy-on-write behavior of pandas (#8940)

* mirror the `Dataset` `copy_coords` test

* avoid modifying the values of a `pandas` series

xarray/tests/test_dataarray.py

@@ -3945,7 +3945,8 @@ def test_copy_coords(self, deep, expected_orig) -> None:
             dims=["a", "b", "c"],
         )
         da_cp = da.copy(deep)
-        da_cp["a"].data[0] = 999
+        new_a = np.array([999, 2])
+        da_cp.coords["a"] = da_cp["a"].copy(data=new_a)
 
         expected_cp = xr.DataArray(
             xr.IndexVariable("a", np.array([999, 2])),

xarray/tests/test_missing.py

@@ -164,9 +164,10 @@ def test_interpolate_pd_compat_non_uniform_index():
             # for the linear methods. This next line inforces the xarray
             # fill_value convention on the pandas output. Therefore, this test
             # only checks that interpolated values are the same (not nans)
-            expected.values[pd.isnull(actual.values)] = np.nan
+            expected_values = expected.values.copy()
+            expected_values[pd.isnull(actual.values)] = np.nan
 
-            np.testing.assert_allclose(actual.values, expected.values)
+            np.testing.assert_allclose(actual.values, expected_values)
 
 
 @requires_scipy