BUG: DataFrame.apply with func altering row in-place

doc/source/whatsnew/v1.1.1.rst

@@ -20,6 +20,7 @@ Fixed regressions
 - Fixed regression in :class:`pandas.core.groupby.RollingGroupby` where column selection was ignored (:issue:`35486`)
 - Fixed regression in :meth:`DataFrame.shift` with ``axis=1`` and heterogeneous dtypes (:issue:`35488`)
 - Fixed regression in ``.groupby(..).rolling(..)`` where a segfault would occur with ``center=True`` and an odd number of values (:issue:`35552`)
+- Fixed regression in :meth:`DataFrame.apply` where functions that altered the input in-place only operated on a single row (:issue:`35462`)
 
 .. ---------------------------------------------------------------------------
 

pandas/core/apply.py

@@ -389,6 +389,8 @@ def series_generator(self):
         blk = mgr.blocks[0]
 
         for (arr, name) in zip(values, self.index):
+            # GH#35462 re-pin mgr in case setitem changed it
+            ser._mgr = mgr
             blk.values = arr
             ser.name = name
             yield ser

pandas/tests/frame/apply/test_frame_apply.py

@@ -1522,3 +1522,22 @@ def test_apply_dtype(self, col):
         expected = df.dtypes
 
         tm.assert_series_equal(result, expected)
+
+
+def test_apply_mutating():
+    # GH#35462 case where applied func pins a new BlockManager to a row
+    df = pd.DataFrame({"a": range(100), "b": range(100, 200)})
+
+    def func(row):
+        mgr = row._mgr
+        row.loc["a"] += 1
+        assert row._mgr is not mgr
+        return row
+
+    expected = df.copy()
+    expected["a"] += 1
+
+    result = df.apply(func, axis=1)
+
+    tm.assert_frame_equal(result, expected)
+    tm.assert_frame_equal(df, result)