ENH: Add lazy copy for sort_values (#50643)

Author: phofl

doc/source/whatsnew/v2.0.0.rst

@@ -952,6 +952,7 @@ Indexing
 - Bug in :meth:`DataFrame.__setitem__` raising when indexer is a :class:`DataFrame` with ``boolean`` dtype (:issue:`47125`)
 - Bug in :meth:`DataFrame.reindex` filling with wrong values when indexing columns and index for ``uint`` dtypes (:issue:`48184`)
 - Bug in :meth:`DataFrame.loc` when setting :class:`DataFrame` with different dtypes coercing values to single dtype (:issue:`50467`)
+- Bug in :meth:`DataFrame.sort_values` where ``None`` was not returned when ``by`` is empty list and ``inplace=True`` (:issue:`50643`)
 - Bug in :meth:`DataFrame.loc` coercing dtypes when setting values with a list indexer (:issue:`49159`)
 - Bug in :meth:`Series.loc` raising error for out of bounds end of slice indexer (:issue:`50161`)
 - Bug in :meth:`DataFrame.loc` raising ``ValueError`` with ``bool`` indexer and :class:`MultiIndex` (:issue:`47687`)

pandas/core/frame.py

@@ -48,6 +48,7 @@
 from pandas._libs.hashtable import duplicated
 from pandas._libs.lib import (
     NoDefault,
+    array_equal_fast,
     no_default,
 )
 from pandas._typing import (
@@ -6695,7 +6696,16 @@ def sort_values(
                 k, kind=kind, ascending=ascending, na_position=na_position, key=key
             )
         else:
-            return self.copy()
+            if inplace:
+                return self._update_inplace(self)
+            else:
+                return self.copy(deep=None)
+
+        if array_equal_fast(indexer, np.arange(0, len(indexer), dtype=indexer.dtype)):
+            if inplace:
+                return self._update_inplace(self)
+            else:
+                return self.copy(deep=None)
 
         new_data = self._mgr.take(
             indexer, axis=self._get_block_manager_axis(axis), verify=False

pandas/core/series.py

@@ -3559,6 +3559,13 @@ def sort_values(
         values_to_sort = ensure_key_mapped(self, key)._values if key else self._values
         sorted_index = nargsort(values_to_sort, kind, bool(ascending), na_position)
 
+        if array_equal_fast(
+            sorted_index, np.arange(0, len(sorted_index), dtype=sorted_index.dtype)
+        ):
+            if inplace:
+                return self._update_inplace(self)
+            return self.copy(deep=None)
+
         result = self._constructor(
             self._values[sorted_index], index=self.index[sorted_index]
         )

pandas/tests/copy_view/test_methods.py

@@ -644,6 +644,45 @@ def test_sort_index(using_copy_on_write):
     tm.assert_series_equal(ser, ser_orig)
 
 
+@pytest.mark.parametrize(
+    "obj, kwargs",
+    [(Series([1, 2, 3], name="a"), {}), (DataFrame({"a": [1, 2, 3]}), {"by": "a"})],
+)
+def test_sort_values(using_copy_on_write, obj, kwargs):
+    obj_orig = obj.copy()
+    obj2 = obj.sort_values(**kwargs)
+
+    if using_copy_on_write:
+        assert np.shares_memory(get_array(obj2, "a"), get_array(obj, "a"))
+    else:
+        assert not np.shares_memory(get_array(obj2, "a"), get_array(obj, "a"))
+
+    # mutating df triggers a copy-on-write for the column / block
+    obj2.iloc[0] = 0
+    assert not np.shares_memory(get_array(obj2, "a"), get_array(obj, "a"))
+    tm.assert_equal(obj, obj_orig)
+
+
+@pytest.mark.parametrize(
+    "obj, kwargs",
+    [(Series([1, 2, 3], name="a"), {}), (DataFrame({"a": [1, 2, 3]}), {"by": "a"})],
+)
+def test_sort_values_inplace(using_copy_on_write, obj, kwargs, using_array_manager):
+    obj_orig = obj.copy()
+    view = obj[:]
+    obj.sort_values(inplace=True, **kwargs)
+
+    assert np.shares_memory(get_array(obj, "a"), get_array(view, "a"))
+
+    # mutating obj triggers a copy-on-write for the column / block
+    obj.iloc[0] = 0
+    if using_copy_on_write:
+        assert not np.shares_memory(get_array(obj, "a"), get_array(view, "a"))
+        tm.assert_equal(view, obj_orig)
+    else:
+        assert np.shares_memory(get_array(obj, "a"), get_array(view, "a"))
+
+
 def test_reorder_levels(using_copy_on_write):
     index = MultiIndex.from_tuples(
         [(1, 1), (1, 2), (2, 1), (2, 2)], names=["one", "two"]

pandas/tests/frame/methods/test_sort_values.py

@@ -619,6 +619,14 @@ def test_sort_values_reshaping(self):
 
         tm.assert_frame_equal(df, expected)
 
+    def test_sort_values_no_by_inplace(self):
+        # GH#50643
+        df = DataFrame({"a": [1, 2, 3]})
+        expected = df.copy()
+        result = df.sort_values(by=[], inplace=True)
+        tm.assert_frame_equal(df, expected)
+        assert result is None
+
 
 class TestDataFrameSortKey:  # test key sorting (issue 27237)
     def test_sort_values_inplace_key(self, sort_by_key):