Support "None" as a variable name and use it as a default

test/__init__.py

@@ -83,7 +83,8 @@ def assertArrayEqual(self, a1, a2):
     def assertDatasetEqual(self, d1, d2):
         # this method is functionally equivalent to `assert d1 == d2`, but it
         # checks each aspect of equality separately for easier debugging
-        self.assertEqual(sorted(d1.variables), sorted(d2.variables))
+        self.assertEqual(sorted(d1.variables, key=str),
+                         sorted(d2.variables, key=str))
         for k in d1:
             v1 = d1.variables[k]
             v2 = d2.variables[k]
@@ -93,14 +94,16 @@ def assertDatasetIdentical(self, d1, d2):
         # this method is functionally equivalent to `assert d1.identical(d2)`,
         # but it checks each aspect of equality separately for easier debugging
         assert utils.dict_equiv(d1.attrs, d2.attrs), (d1.attrs, d2.attrs)
-        self.assertEqual(sorted(d1.variables), sorted(d2.variables))
+        self.assertEqual(sorted(d1.variables, key=str),
+                         sorted(d2.variables, key=str))
         for k in d1:
             v1 = d1.variables[k]
             v2 = d2.variables[k]
             assert v1.identical(v2), (v1, v2)
 
     def assertDatasetAllClose(self, d1, d2, rtol=1e-05, atol=1e-08):
-        self.assertEqual(sorted(d1.variables), sorted(d2.variables))
+        self.assertEqual(sorted(d1.variables, key=str),
+                         sorted(d2.variables, key=str))
         for k in d1:
             v1 = d1.variables[k]
             v2 = d2.variables[k]

test/test_backends.py

@@ -76,6 +76,11 @@ def test_roundtrip_test_data(self):
         actual = self.roundtrip(expected)
         self.assertDatasetAllClose(expected, actual)
 
+    def test_roundtrip_None_variable(self):
+        expected = Dataset({None: (('x', 'y'), [[0, 1], [2, 3]])})
+        actual = self.roundtrip(expected)
+        self.assertDatasetAllClose(expected, actual)
+
     def test_roundtrip_string_data(self):
         expected = Dataset({'x': ('t', ['abc', 'def'])})
         actual = self.roundtrip(expected)

test/test_data_array.py

@@ -229,6 +229,21 @@ def test_dataset_math(self):
                             'x': sim['x']})['tmin']
         self.assertDataArrayEqual(actual, expected)
 
+    def test_math_name(self):
+        # Verify that name is preserved only when it can be done unambiguously.
+        # The rule (copied from pandas.Series) is keep the current name only if
+        # the other object has no name attribute and this object isn't a
+        # coordinate; otherwise reset to None.
+        ds = self.ds
+        a = self.dv
+        self.assertEqual((+a).name, 'foo')
+        self.assertEqual((a + 0).name, 'foo')
+        self.assertIs((a + a.rename(None)).name, None)
+        self.assertIs((a + a).name, None)
+        self.assertIs((+ds['x']).name, None)
+        self.assertIs((ds['x'] + 0).name, None)
+        self.assertIs((a + ds['x']).name, None)
+
     def test_coord_math(self):
         ds = Dataset({'x': ('x', 1 + np.arange(3))})
         expected = ds.copy()
@@ -381,3 +396,8 @@ def test_to_and_from_series(self):
         self.assertEqual('foo', actual.name)
         # test roundtrip
         self.assertDataArrayIdentical(self.dv, DataArray.from_series(actual))
+        # test name is None
+        actual.name = None
+        expected_da = self.dv.rename(None)
+        self.assertDataArrayIdentical(expected_da,
+                                      DataArray.from_series(actual))

test/test_dataset.py

@@ -171,6 +171,10 @@ def test_equals_and_identical(self):
         self.assertFalse(data.equals(data2))
         self.assertTrue(data != data2)
 
+        data = create_test_data(seed=42).rename({'var1': None})
+        self.assertTrue(data.equals(data))
+        self.assertTrue(data.identical(data))
+
     def test_attrs(self):
         data = create_test_data(seed=42)
         data.attrs = {'foobar': 'baz'}

xray/backends/common.py

@@ -2,6 +2,21 @@
 from xray.pycompat import iteritems
 
 
+NONE_VAR_NAME = '__values__'
+
+
+def _encode_variable_name(name):
+    if name is None:
+        name = NONE_VAR_NAME
+    return name
+
+
+def _decode_variable_name(name):
+    if name == NONE_VAR_NAME:
+        name = None
+    return name
+
+
 class AbstractDataStore(object):
     def open_store_variable(self, v):
         raise NotImplementedError
@@ -12,7 +27,8 @@ def store_variables(self):
 
     @property
     def variables(self):
-        return FrozenOrderedDict((k, self.open_store_variable(v))
+        return FrozenOrderedDict((_decode_variable_name(k),
+                                  self.open_store_variable(v))
                                  for k, v in iteritems(self.store_variables))
 
     def sync(self):
@@ -39,7 +55,7 @@ def set_attributes(self, attributes):
 
     def set_variables(self, variables):
         for vn, v in iteritems(variables):
-            self.set_variable(vn, v)
+            self.set_variable(_encode_variable_name(vn), v)
 
     def set_necessary_dimensions(self, variable):
         for d, l in zip(variable.dimensions, variable.shape):

xray/common.py

@@ -131,7 +131,10 @@ def _wrap_indent(text, start='', length=None):
 
 
 def array_repr(arr):
-    name_str = ('%r ' % arr.name) if hasattr(arr, 'name') else ''
+    if hasattr(arr, 'name') and arr.name is not None:
+        name_str = '%r ' % arr.name
+    else:
+        name_str = ''
     dim_summary = ', '.join('%s: %s' % (k, v) for k, v
                             in zip(arr.dimensions, arr.shape))
     summary = ['<xray.%s %s(%s)>'% (type(arr).__name__, name_str, dim_summary)]

xray/data_array.py

@@ -345,7 +345,8 @@ def rename(self, new_name_or_name_dict):
         --------
         Dataset.rename
         """
-        if isinstance(new_name_or_name_dict, basestring):
+        if (isinstance(new_name_or_name_dict, basestring)
+                or new_name_or_name_dict is None):
             new_name = new_name_or_name_dict
             name_dict = {self.name: new_name}
         else:
@@ -597,10 +598,9 @@ def from_series(cls, series):
         with NaN). Thus this operation should be the inverse of the `to_series`
         method.
         """
-        name = series.name if series.name is not None else 'values'
-        df = pd.DataFrame({name: series})
+        df = pd.DataFrame({series.name: series})
         ds = xray.Dataset.from_dataframe(df)
-        return ds[name]
+        return ds[series.name]
 
     def equals(self, other):
         """True if two DataArrays have the same dimensions, coordinates and
@@ -638,29 +638,22 @@ def identical(self, other):
     def _select_coordinates(self):
         return xray.Dataset(self.coordinates)
 
-    def _refocus(self, new_var, name=None):
-        """Returns a copy of this DataArray's dataset with this
-        DataArray's focus variable replaced by `new_var`.
-
-        If `new_var` is a DataArray, its contents will be merged in.
-        """
-        if not hasattr(new_var, 'dimensions'):
-            new_var = variable.Variable(self.variable.dimensions, new_var)
+    def __array_wrap__(self, obj, context=None):
+        new_var = self.variable.__array_wrap__(obj, context)
         ds = self._select_coordinates()
-        if name is None:
-            name = self.name + '_'
+        if (self.name,) == self.dimensions:
+            # use a new name for coordinate variables
+            name = None
+        else:
+            name = self.name
         ds[name] = new_var
         return ds[name]
 
-    def __array_wrap__(self, obj, context=None):
-        return self._refocus(self.variable.__array_wrap__(obj, context))
-
     @staticmethod
     def _unary_op(f):
         @functools.wraps(f)
         def func(self, *args, **kwargs):
-            return self._refocus(f(self.variable, *args, **kwargs),
-                                 self.name + '_' + f.__name__)
+            return self.__array_wrap__(f(self.variable, *args, **kwargs))
         return func
 
     def _check_coordinates_compat(self, other):
@@ -682,11 +675,13 @@ def func(self, other):
             if hasattr(other, 'coordinates'):
                 ds.merge(other.coordinates, inplace=True)
             other_array = getattr(other, 'variable', other)
-            other_name = getattr(other, 'name', 'other')
-            name = self.name + '_' + f.__name__ + '_' + other_name
+            if hasattr(other, 'name') or (self.name,) == self.dimensions:
+                name = None
+            else:
+                name = self.name
             ds[name] = (f(self.variable, other_array)
-                         if not reflexive
-                         else f(other_array, self.variable))
+                        if not reflexive
+                        else f(other_array, self.variable))
             return ds[name]
         return func
 

xray/dataset.py

@@ -231,6 +231,14 @@ def _assert_compat_valid(compat):
                          "'identical'" % compat)
 
 
+def _item0_str(items):
+    """Key function for use in sorted on a list of variables.
+
+    This is useful because None is not comparable to strings in Python 3.
+    """
+    return str(items[0])
+
+
 def as_dataset(obj):
     """Cast the given object to a Dataset.
 
@@ -469,8 +477,10 @@ def equals(self, other):
             return (len(self) == len(other)
                     and all(k1 == k2 and v1.equals(v2)
                             for (k1, v1), (k2, v2)
-                            in zip(sorted(self.variables.items()),
-                                   sorted(other.variables.items()))))
+                            in zip(sorted(self.variables.items(),
+                                          key=_item0_str),
+                                   sorted(other.variables.items(),
+                                          key=_item0_str))))
         except (TypeError, AttributeError):
             return False
 
@@ -489,8 +499,10 @@ def identical(self, other):
                     and len(self) == len(other)
                     and all(k1 == k2 and v1.identical(v2)
                             for (k1, v1), (k2, v2)
-                            in zip(sorted(self.variables.items()),
-                                   sorted(other.variables.items()))))
+                            in zip(sorted(self.variables.items(),
+                                          key=_item0_str),
+                                   sorted(other.variables.items(),
+                                          key=_item0_str))))
         except (TypeError, AttributeError):
             return False