REGR: DataFrame.apply() with raw option and func returning string (#36610)

Author: simonjayhawkins

doc/source/whatsnew/v1.1.3.rst

@@ -35,6 +35,7 @@ Fixed regressions
 - Fixed regression in :meth:`Series.__getitem__` incorrectly raising when the input was a frozenset (:issue:`35747`)
 - Fixed regression in :meth:`read_excel` with ``engine="odf"`` caused ``UnboundLocalError`` in some cases where cells had nested child nodes (:issue:`36122`, :issue:`35802`)
 - Fixed regression in :class:`DataFrame` and :class:`Series` comparisons between numeric arrays and strings (:issue:`35700`, :issue:`36377`)
+- Fixed regression in :meth:`DataFrame.apply` with ``raw=True`` and user-function returning string (:issue:`35940`)
 - Fixed regression when setting empty :class:`DataFrame` column to a :class:`Series` in preserving name of index in frame (:issue:`36527`)
 - Fixed regression in :class:`Period` incorrect value for ordinal over the maximum timestamp (:issue:`36430`)
 

pandas/core/apply.py

@@ -216,7 +216,23 @@ def apply_empty_result(self):
 
     def apply_raw(self):
         """ apply to the values as a numpy array """
-        result = np.apply_along_axis(self.f, self.axis, self.values)
+
+        def wrap_function(func):
+            """
+            Wrap user supplied function to work around numpy issue.
+
+            see https://github.com/numpy/numpy/issues/8352
+            """
+
+            def wrapper(*args, **kwargs):
+                result = func(*args, **kwargs)
+                if isinstance(result, str):
+                    result = np.array(result, dtype=object)
+                return result
+
+            return wrapper
+
+        result = np.apply_along_axis(wrap_function(self.f), self.axis, self.values)
 
         # TODO: mixed type case
         if result.ndim == 2:

pandas/tests/frame/apply/test_frame_apply.py

@@ -1545,3 +1545,11 @@ def test_apply_no_suffix_index():
     )
 
     tm.assert_frame_equal(result, expected)
+
+
+def test_apply_raw_returns_string():
+    # https://github.com/pandas-dev/pandas/issues/35940
+    df = pd.DataFrame({"A": ["aa", "bbb"]})
+    result = df.apply(lambda x: x[0], axis=1, raw=True)
+    expected = pd.Series(["aa", "bbb"])
+    tm.assert_series_equal(result, expected)