Fix pandas < v0.15.2 and GH700

Author: shoyer

xray/backends/api.py

@@ -170,9 +170,8 @@ def maybe_decode_store(store, lock=False):
             else:
                 file_arg = filename_or_obj
             token = tokenize(file_arg, group, decode_cf, mask_and_scale,
-                             decode_times, concat_characters,
-                             decode_coords, engine, chunks, lock,
-                             drop_variables)
+                             decode_times, concat_characters, decode_coords,
+                             engine, chunks, drop_variables)
             name_prefix = '%s:%s/' % (filename_or_obj, group or '')
             ds2 = ds.chunk(chunks, name_prefix=name_prefix, token=token,
                            lock=lock)

xray/core/indexing.py

@@ -386,9 +386,11 @@ class PandasIndexAdapter(utils.NDArrayMixin):
     def __init__(self, array, dtype=None):
         self.array = utils.safe_cast_to_index(array)
         if dtype is None:
-            # if a PeriodIndex, force an object dtype
             if isinstance(array, pd.PeriodIndex):
                 dtype = np.dtype('O')
+            elif hasattr(array, 'categories'):
+                # category isn't a real numpy dtype
+                dtype = array.categories.dtype
             else:
                 dtype = array.dtype
         self._dtype = dtype
@@ -405,7 +407,12 @@ def __array__(self, dtype=None):
             with suppress(AttributeError):
                 # this might not be public API
                 array = array.asobject
-        return array.values.astype(dtype)
+        return np.asarray(array.values, dtype=dtype)
+
+    @property
+    def shape(self):
+        # .shape is broken on pandas prior to v0.15.2
+        return (len(self.array),)
 
     def __getitem__(self, key):
         if isinstance(key, tuple) and len(key) == 1:

xray/test/test_dataarray.py

@@ -1413,6 +1413,15 @@ def test_to_and_from_series(self):
         self.assertDataArrayIdentical(expected_da,
                                       DataArray.from_series(actual))
 
+    def test_series_categorical_index(self):
+        # regression test for GH700
+        if not hasattr(pd, 'CategoricalIndex'):
+            raise unittest.SkipTest('requires pandas with CategoricalIndex')
+
+        s = pd.Series(range(5), index=pd.CategoricalIndex(list('aabbc')))
+        arr = DataArray(s)
+        assert "'a'" in repr(arr)  # should not error
+
     def test_to_masked_array(self):
         rs = np.random.RandomState(44)
         x = rs.random_sample(size=(10, 20))

xray/test/test_variable.py

@@ -163,10 +163,8 @@ def test_datetime64_conversion(self):
             self.assertArrayEqual(v.values, times.values)
             self.assertEqual(v.values.dtype, np.dtype('datetime64[ns]'))
             same_source = source_ndarray(v.values) is source_ndarray(values)
-            if preserve_source and self.cls is Variable:
+            if preserve_source:
                 self.assertTrue(same_source)
-            else:
-                self.assertFalse(same_source)
 
     def test_timedelta64_conversion(self):
         times = pd.timedelta_range(start=0, periods=3)
@@ -181,10 +179,8 @@ def test_timedelta64_conversion(self):
             self.assertArrayEqual(v.values, times.values)
             self.assertEqual(v.values.dtype, np.dtype('timedelta64[ns]'))
             same_source = source_ndarray(v.values) is source_ndarray(values)
-            if preserve_source and self.cls is Variable:
+            if preserve_source:
                 self.assertTrue(same_source)
-            else:
-                self.assertFalse(same_source)
 
     def test_object_conversion(self):
         data = np.arange(5).astype(str).astype(object)