DEPR: Index.reindex with duplicate index (#42568)

Author: jbrockmendel

doc/source/user_guide/duplicates.rst

@@ -28,6 +28,7 @@ duplicates present. The output can't be determined, and so pandas raises.
 
 .. ipython:: python
    :okexcept:
+   :okwarning:
 
    s1 = pd.Series([0, 1, 2], index=["a", "b", "b"])
    s1.reindex(["a", "b", "c"])

doc/source/whatsnew/v1.4.0.rst

@@ -162,6 +162,7 @@ Deprecations
 - Deprecated ignoring missing labels when indexing with a sequence of labels on a level of a MultiIndex (:issue:`42351`)
 - Creating an empty Series without a dtype will now raise a more visible ``FutureWarning`` instead of a ``DeprecationWarning`` (:issue:`30017`)
 - Deprecated the 'kind' argument in :meth:`Index.get_slice_bound`, :meth:`Index.slice_indexer`, :meth:`Index.slice_locs`; in a future version passing 'kind' will raise (:issue:`42857`)
+- Deprecated :meth:`Index.reindex` with a non-unique index (:issue:`42568`)
 -
 
 .. ---------------------------------------------------------------------------

pandas/core/indexes/base.py

@@ -3915,6 +3915,15 @@ def reindex(
                         )
                     indexer, _ = self.get_indexer_non_unique(target)
 
+                if not self.is_unique:
+                    # GH#42568
+                    warnings.warn(
+                        "reindexing with a non-unique Index is deprecated and "
+                        "will raise in a future version",
+                        FutureWarning,
+                        stacklevel=2,
+                    )
+
         target = self._wrap_reindex_result(target, indexer, preserve_names)
         return target, indexer
 

pandas/core/indexes/category.py

@@ -426,6 +426,14 @@ def reindex(
             missing = np.array([], dtype=np.intp)
         else:
             indexer, missing = self.get_indexer_non_unique(target)
+            if not self.is_unique:
+                # GH#42568
+                warnings.warn(
+                    "reindexing with a non-unique Index is deprecated and will "
+                    "raise in a future version",
+                    FutureWarning,
+                    stacklevel=2,
+                )
 
         if len(self) and indexer is not None:
             new_target = self.take(indexer)

pandas/tests/frame/indexing/test_getitem.py

@@ -301,7 +301,8 @@ def test_getitem_boolean_frame_unaligned_with_duplicate_columns(self, df_dup_col
         df = df_dup_cols
         msg = "cannot reindex on an axis with duplicate labels"
         with pytest.raises(ValueError, match=msg):
-            df[df.A > 6]
+            with tm.assert_produces_warning(FutureWarning, match="non-unique"):
+                df[df.A > 6]
 
     def test_getitem_boolean_series_with_duplicate_columns(self, df_dup_cols):
         # boolean indexing

pandas/tests/frame/indexing/test_setitem.py

@@ -70,7 +70,8 @@ def test_setitem_error_msmgs(self):
         )
         msg = "cannot reindex on an axis with duplicate labels"
         with pytest.raises(ValueError, match=msg):
-            df["newcol"] = ser
+            with tm.assert_produces_warning(FutureWarning, match="non-unique"):
+                df["newcol"] = ser
 
         # GH 4107, more descriptive error message
         df = DataFrame(np.random.randint(0, 2, (4, 4)), columns=["a", "b", "c", "d"])

pandas/tests/frame/methods/test_reindex.py

@@ -674,7 +674,8 @@ def test_reindex_dups(self):
         # reindex fails
         msg = "cannot reindex on an axis with duplicate labels"
         with pytest.raises(ValueError, match=msg):
-            df.reindex(index=list(range(len(df))))
+            with tm.assert_produces_warning(FutureWarning, match="non-unique"):
+                df.reindex(index=list(range(len(df))))
 
     def test_reindex_with_duplicate_columns(self):
 
@@ -684,9 +685,11 @@ def test_reindex_with_duplicate_columns(self):
         )
         msg = "cannot reindex on an axis with duplicate labels"
         with pytest.raises(ValueError, match=msg):
-            df.reindex(columns=["bar"])
+            with tm.assert_produces_warning(FutureWarning, match="non-unique"):
+                df.reindex(columns=["bar"])
         with pytest.raises(ValueError, match=msg):
-            df.reindex(columns=["bar", "foo"])
+            with tm.assert_produces_warning(FutureWarning, match="non-unique"):
+                df.reindex(columns=["bar", "foo"])
 
     def test_reindex_axis_style(self):
         # https://github.com/pandas-dev/pandas/issues/12392
@@ -958,7 +961,8 @@ def test_reindex_with_categoricalindex(self):
         # passed duplicate indexers are not allowed
         msg = "cannot reindex on an axis with duplicate labels"
         with pytest.raises(ValueError, match=msg):
-            df2.reindex(["a", "b"])
+            with tm.assert_produces_warning(FutureWarning, match="non-unique"):
+                df2.reindex(["a", "b"])
 
         # args NotImplemented ATM
         msg = r"argument {} is not implemented for CategoricalIndex\.reindex"

pandas/tests/indexes/categorical/test_reindex.py

@@ -13,26 +13,32 @@
 
 class TestReindex:
     def test_reindex_dtype(self):
-        c = CategoricalIndex(["a", "b", "c", "a"])
-        res, indexer = c.reindex(["a", "c"])
+        # GH#11586
+        ci = CategoricalIndex(["a", "b", "c", "a"])
+        with tm.assert_produces_warning(FutureWarning, match="non-unique"):
+            res, indexer = ci.reindex(["a", "c"])
+
         tm.assert_index_equal(res, Index(["a", "a", "c"]), exact=True)
         tm.assert_numpy_array_equal(indexer, np.array([0, 3, 2], dtype=np.intp))
 
-        c = CategoricalIndex(["a", "b", "c", "a"])
-        res, indexer = c.reindex(Categorical(["a", "c"]))
+        ci = CategoricalIndex(["a", "b", "c", "a"])
+        with tm.assert_produces_warning(FutureWarning, match="non-unique"):
+            res, indexer = ci.reindex(Categorical(["a", "c"]))
 
         exp = CategoricalIndex(["a", "a", "c"], categories=["a", "c"])
         tm.assert_index_equal(res, exp, exact=True)
         tm.assert_numpy_array_equal(indexer, np.array([0, 3, 2], dtype=np.intp))
 
-        c = CategoricalIndex(["a", "b", "c", "a"], categories=["a", "b", "c", "d"])
-        res, indexer = c.reindex(["a", "c"])
+        ci = CategoricalIndex(["a", "b", "c", "a"], categories=["a", "b", "c", "d"])
+        with tm.assert_produces_warning(FutureWarning, match="non-unique"):
+            res, indexer = ci.reindex(["a", "c"])
         exp = Index(["a", "a", "c"], dtype="object")
         tm.assert_index_equal(res, exp, exact=True)
         tm.assert_numpy_array_equal(indexer, np.array([0, 3, 2], dtype=np.intp))
 
-        c = CategoricalIndex(["a", "b", "c", "a"], categories=["a", "b", "c", "d"])
-        res, indexer = c.reindex(Categorical(["a", "c"]))
+        ci = CategoricalIndex(["a", "b", "c", "a"], categories=["a", "b", "c", "d"])
+        with tm.assert_produces_warning(FutureWarning, match="non-unique"):
+            res, indexer = ci.reindex(Categorical(["a", "c"]))
         exp = CategoricalIndex(["a", "a", "c"], categories=["a", "c"])
         tm.assert_index_equal(res, exp, exact=True)
         tm.assert_numpy_array_equal(indexer, np.array([0, 3, 2], dtype=np.intp))

pandas/tests/indexes/multi/test_reindex.py

@@ -110,7 +110,8 @@ def test_reindex_non_unique():
 
     msg = "cannot handle a non-unique multi-index!"
     with pytest.raises(ValueError, match=msg):
-        a.reindex(new_idx)
+        with tm.assert_produces_warning(FutureWarning, match="non-unique"):
+            a.reindex(new_idx)
 
 
 @pytest.mark.parametrize("values", [[["a"], ["x"]], [[], []]])

pandas/tests/resample/test_datetime_index.py

@@ -694,7 +694,8 @@ def test_asfreq_non_unique():
 
     msg = "cannot reindex on an axis with duplicate labels"
     with pytest.raises(ValueError, match=msg):
-        ts.asfreq("B")
+        with tm.assert_produces_warning(FutureWarning, match="non-unique"):
+            ts.asfreq("B")
 
 
 def test_resample_axis1():