[#16737] Index type for Series with empty data

doc/source/whatsnew/v1.1.0.rst

@@ -56,6 +56,7 @@ Other API changes
 - :meth:`Series.describe` will now show distribution percentiles for ``datetime`` dtypes, statistics ``first`` and ``last``
   will now be ``min`` and ``max`` to match with numeric dtypes in :meth:`DataFrame.describe` (:issue:`30164`)
 - :meth:`Groupby.groups` now returns an abbreviated representation when called on large dataframes (:issue:`1135`)
+- :class:`Series` constructor will default to construct an :class:`Index`, rather than an :class:`RangeIndex` when constructed with empty data, matching the behaviour of ``data=None``.
 
 Backwards incompatible API changes
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

pandas/core/indexes/base.py

@@ -5563,7 +5563,9 @@ def _validate_join_method(method):
 def default_index(n):
     from pandas.core.indexes.range import RangeIndex
 
-    return RangeIndex(0, n, name=None)
+    if n == 0:
+        return Index([])
+    return RangeIndex(0, n)
 
 
 def maybe_extract_name(name, obj, cls) -> Optional[Hashable]:

pandas/tests/extension/base/missing.py

@@ -19,7 +19,7 @@ def test_isna(self, data_missing):
 
         # GH 21189
         result = pd.Series(data_missing).drop([0, 1]).isna()
-        expected = pd.Series([], dtype=bool)
+        expected = pd.Series([], dtype=bool, index=pd.RangeIndex(0))
         self.assert_series_equal(result, expected)
 
     def test_dropna_array(self, data_missing):

pandas/tests/extension/decimal/test_decimal.py

@@ -86,7 +86,7 @@ def convert(x):
         else:
             right_na = right.isna()
 
-        tm.assert_series_equal(left_na, right_na)
+        tm.assert_series_equal(left_na, right_na, *args, **kwargs)
         return tm.assert_series_equal(left[~left_na], right[~right_na], *args, **kwargs)
 
     @classmethod

pandas/tests/extension/test_sparse.py

@@ -185,7 +185,7 @@ def test_isna(self, data_missing):
 
         # GH 21189
         result = pd.Series(data_missing).drop([0, 1]).isna()
-        expected = pd.Series([], dtype=expected_dtype)
+        expected = pd.Series([], dtype=expected_dtype, index=pd.RangeIndex(0))
         self.assert_series_equal(result, expected)
 
     def test_fillna_limit_pad(self, data_missing):

pandas/tests/frame/test_constructors.py

@@ -84,9 +84,9 @@ def test_empty_constructor(self, constructor):
     @pytest.mark.parametrize(
         "emptylike,expected_index,expected_columns",
         [
-            ([[]], RangeIndex(1), RangeIndex(0)),
-            ([[], []], RangeIndex(2), RangeIndex(0)),
-            ([(_ for _ in [])], RangeIndex(1), RangeIndex(0)),
+            ([[]], RangeIndex(1), Index([])),
+            ([[], []], RangeIndex(2), Index([])),
+            ([(_ for _ in [])], RangeIndex(1), Index([])),
         ],
     )
     def test_emptylike_constructor(self, emptylike, expected_index, expected_columns):
@@ -337,7 +337,7 @@ def test_constructor_dict(self):
 
         # with dict of empty list and Series
         frame = DataFrame({"A": [], "B": []}, columns=["A", "B"])
-        tm.assert_index_equal(frame.index, Index([], dtype=np.int64))
+        tm.assert_index_equal(frame.index, Index([]))
 
         # GH 14381
         # Dict with None value

pandas/tests/groupby/test_filters.py

@@ -97,7 +97,7 @@ def test_filter_out_all_groups_in_df():
     df = pd.DataFrame({"a": [1, 1, 2], "b": [1, 2, 0]})
     res = df.groupby("a")
     res = res.filter(lambda x: x["b"].sum() > 5, dropna=True)
-    expected = pd.DataFrame({"a": [], "b": []}, dtype="int64")
+    expected = pd.DataFrame({"a": [], "b": []}, dtype="int64", index=pd.RangeIndex(0))
     tm.assert_frame_equal(expected, res)
 
 

pandas/tests/groupby/test_grouping.py

@@ -611,10 +611,7 @@ def test_list_grouper_with_nat(self):
     @pytest.mark.parametrize(
         "func,expected",
         [
-            (
-                "transform",
-                pd.Series(name=2, dtype=np.float64, index=pd.RangeIndex(0, 0, 1)),
-            ),
+            ("transform", pd.Series(name=2, dtype=np.float64)),
             (
                 "agg",
                 pd.Series(name=2, dtype=np.float64, index=pd.Float64Index([], name=1)),
@@ -638,10 +635,13 @@ def test_evaluate_with_empty_groups(self, func, expected):
     def test_groupby_empty(self):
         # https://github.com/pandas-dev/pandas/issues/27190
         s = pd.Series([], name="name", dtype="float64")
-        gr = s.groupby([])
 
+        expected = s.copy()
+        expected.index = pd.RangeIndex(0)
+
+        gr = s.groupby([])
         result = gr.mean()
-        tm.assert_series_equal(result, s)
+        tm.assert_series_equal(result, expected)
 
         # check group properties
         assert len(gr.grouper.groupings) == 1

pandas/tests/io/json/test_pandas.py

@@ -255,13 +255,16 @@ def test_roundtrip_empty(self, orient, convert_axes, numpy):
         )
         expected = self.empty_frame.copy()
 
-        # TODO: both conditions below are probably bugs
+        # TODO: conditions below are probably bugs
         if convert_axes:
-            expected.index = expected.index.astype(float)
             expected.columns = expected.columns.astype(float)
+
         if numpy and orient == "values":
             expected = expected.reindex([0], axis=1).reset_index(drop=True)
 
+        if convert_axes:
+            expected.index = expected.index.astype(float)
+
         tm.assert_frame_equal(result, expected)
 
     @pytest.mark.parametrize("convert_axes", [True, False])
@@ -675,8 +678,7 @@ def test_series_roundtrip_empty(self, orient, numpy):
         # TODO: see what causes inconsistency
         if orient in ("values", "records"):
             expected = expected.reset_index(drop=True)
-        else:
-            expected.index = expected.index.astype(float)
+        expected.index = expected.index.astype(float)
 
         tm.assert_series_equal(result, expected)
 

pandas/tests/reshape/test_concat.py

@@ -2199,7 +2199,8 @@ def test_concat_empty_series_timelike(self, tz, values):
             {
                 0: pd.Series([pd.NaT] * len(values), dtype="M8[ns]").dt.tz_localize(tz),
                 1: values,
-            }
+            },
+            index=pd.Index(list(range(len(values))), dtype="object"),
         )
         result = concat([first, second], axis=1)
         tm.assert_frame_equal(result, expected)

pandas/tests/series/methods/test_interpolate.py

@@ -169,7 +169,7 @@ def test_interpolate_corners(self, kwargs):
         s = Series([np.nan, np.nan])
         tm.assert_series_equal(s.interpolate(**kwargs), s)
 
-        s = Series([], dtype=object).interpolate()
+        s = Series([], dtype=object, index=pd.RangeIndex(0)).interpolate()
         tm.assert_series_equal(s.interpolate(**kwargs), s)
 
     def test_interpolate_index_values(self):

pandas/tests/series/test_constructors.py

@@ -127,37 +127,40 @@ def test_constructor_empty(self, input_class):
         with tm.assert_produces_warning(DeprecationWarning, check_stacklevel=False):
             empty = Series()
             empty2 = Series(input_class())
-
-        # these are Index() and RangeIndex() which don't compare type equal
-        # but are just .equals
-        tm.assert_series_equal(empty, empty2, check_index_type=False)
+        tm.assert_series_equal(empty, empty2)
+        assert type(empty.index) is Index
 
         # With explicit dtype:
         empty = Series(dtype="float64")
         empty2 = Series(input_class(), dtype="float64")
-        tm.assert_series_equal(empty, empty2, check_index_type=False)
+        tm.assert_series_equal(empty, empty2)
+        assert type(empty.index) is Index
 
         # GH 18515 : with dtype=category:
         empty = Series(dtype="category")
         empty2 = Series(input_class(), dtype="category")
-        tm.assert_series_equal(empty, empty2, check_index_type=False)
+        tm.assert_series_equal(empty, empty2)
+        assert type(empty.index) is Index
 
         if input_class is not list:
             # With index:
             with tm.assert_produces_warning(DeprecationWarning, check_stacklevel=False):
                 empty = Series(index=range(10))
                 empty2 = Series(input_class(), index=range(10))
             tm.assert_series_equal(empty, empty2)
+            assert type(empty.index) is pd.RangeIndex
 
             # With index and dtype float64:
             empty = Series(np.nan, index=range(10))
             empty2 = Series(input_class(), index=range(10), dtype="float64")
             tm.assert_series_equal(empty, empty2)
+            assert type(empty.index) is pd.RangeIndex
 
             # GH 19853 : with empty string, index and dtype str
             empty = Series("", dtype=str, index=range(3))
             empty2 = Series("", index=range(3))
             tm.assert_series_equal(empty, empty2)
+            assert type(empty.index) is pd.RangeIndex
 
     @pytest.mark.parametrize("input_arg", [np.nan, float("nan")])
     def test_constructor_nan(self, input_arg):