BUG: Series.replace with CoW when made from an Index

doc/source/whatsnew/v3.0.0.rst

@@ -940,6 +940,7 @@ Other
 - Bug in Dataframe Interchange Protocol implementation was returning incorrect results for data buffers' associated dtype, for string and datetime columns (:issue:`54781`)
 - Bug in ``Series.list`` methods not preserving the original :class:`Index`. (:issue:`58425`)
 - Bug in ``Series.list`` methods not preserving the original name. (:issue:`60522`)
+- Bug in ``Series.replace`` when the Series was created from an :class:`Index` and Copy-On-Write is enabled (:issue:`61622`)
 - Bug in printing a :class:`DataFrame` with a :class:`DataFrame` stored in :attr:`DataFrame.attrs` raised a ``ValueError`` (:issue:`60455`)
 - Bug in printing a :class:`Series` with a :class:`DataFrame` stored in :attr:`Series.attrs` raised a ``ValueError`` (:issue:`60568`)
 - Fixed bug where the :class:`DataFrame` constructor misclassified array-like objects with a ``.name`` attribute as :class:`Series` or :class:`Index` (:issue:`61443`)

pandas/core/internals/blocks.py

@@ -10,7 +10,6 @@
     final,
 )
 import warnings
-import weakref
 
 import numpy as np
 
@@ -863,14 +862,22 @@ def replace_list(
                 )
 
                 if i != src_len:
-                    # This is ugly, but we have to get rid of intermediate refs
-                    # that did not go out of scope yet, otherwise we will trigger
-                    # many unnecessary copies
+                    # This is ugly, but we have to get rid of intermediate refs. We
+                    # can simply clear the referenced_blocks if we already copied,
+                    # otherwise we have to remove ourselves
+                    self_blk_ids = {
+                        id(b()): i for i, b in enumerate(self.refs.referenced_blocks)
+                    }
                     for b in result:
-                        ref = weakref.ref(b)
-                        b.refs.referenced_blocks.pop(
-                            b.refs.referenced_blocks.index(ref)
-                        )
+                        if b.refs is self.refs:
+                            # We are still sharing memory with self
+                            if id(b) in self_blk_ids:
+                                # Remove ourselves from the refs; we are temporary
+                                self.refs.referenced_blocks.pop(self_blk_ids[id(b)])
+                        else:
+                            # We have already copied, so we can clear the refs to avoid
+                            # future copies
+                            b.refs.referenced_blocks.clear()
                 new_rb.extend(result)
             rb = new_rb
         return rb

pandas/tests/series/methods/test_replace.py

@@ -715,3 +715,11 @@ def test_replace_all_NA(self):
         result = df.replace({r"^#": "$"}, regex=True)
         expected = pd.Series([pd.NA, pd.NA])
         tm.assert_series_equal(result, expected)
+
+
+def test_replace_from_index():
+    # https://github.com/pandas-dev/pandas/issues/61622
+    idx = pd.Index(["a", "b", "c"], dtype="string[pyarrow]")
+    expected = pd.Series(["d", "b", "c"], dtype="string[pyarrow]")
+    result = pd.Series(idx).replace({"z": "b", "a": "d"})
+    tm.assert_series_equal(result, expected)