BUG: Fix TypeError raised in libreduction (pandas-dev#28643)

Author: jbrockmendel

pandas/_libs/reduction.pyx

@@ -15,15 +15,15 @@ from numpy cimport (ndarray,
 cnp.import_array()
 
 cimport pandas._libs.util as util
-from pandas._libs.lib import maybe_convert_objects, values_from_object
+from pandas._libs.lib import maybe_convert_objects
 
 
 cdef _get_result_array(object obj, Py_ssize_t size, Py_ssize_t cnt):
 
     if (util.is_array(obj) or
             (isinstance(obj, list) and len(obj) == cnt) or
             getattr(obj, 'shape', None) == (cnt,)):
-        raise ValueError('function does not reduce')
+        raise ValueError('Function does not reduce')
 
     return np.empty(size, dtype='O')
 
@@ -103,7 +103,7 @@ cdef class Reducer:
             ndarray arr, result, chunk
             Py_ssize_t i, incr
             flatiter it
-            bint has_labels
+            bint has_labels, has_ndarray_labels
             object res, name, labels, index
             object cached_typ=None
 
@@ -113,14 +113,18 @@ cdef class Reducer:
         chunk.data = arr.data
         labels = self.labels
         has_labels = labels is not None
+        has_ndarray_labels = util.is_array(labels)
         has_index = self.index is not None
         incr = self.increment
 
         try:
             for i in range(self.nresults):
 
-                if has_labels:
+                if has_ndarray_labels:
                     name = util.get_value_at(labels, i)
+                elif has_labels:
+                    # labels is an ExtensionArray
+                    name = labels[i]
                 else:
                     name = None
 
@@ -362,7 +366,8 @@ cdef class SeriesGrouper:
 
     def get_result(self):
         cdef:
-            ndarray arr, result
+            # Define result to avoid UnboundLocalError
+            ndarray arr, result = None
             ndarray[int64_t] labels, counts
             Py_ssize_t i, n, group_size, lab
             object res
@@ -428,6 +433,9 @@ cdef class SeriesGrouper:
             islider.reset()
             vslider.reset()
 
+        if result is None:
+            raise ValueError("No result.")
+
         if result.dtype == np.object_:
             result = maybe_convert_objects(result)
 
@@ -639,11 +647,11 @@ def compute_reduction(arr, f, axis=0, dummy=None, labels=None):
     """
 
     if labels is not None:
-        if labels._has_complex_internals:
-            raise Exception('Cannot use shortcut')
+        # Caller is responsible for ensuring we don't have MultiIndex
+        assert not labels._has_complex_internals
 
-        # pass as an ndarray
-        labels = values_from_object(labels)
+        # pass as an ndarray/ExtensionArray
+        labels = labels._values
 
     reducer = Reducer(arr, f, axis=axis, dummy=dummy, labels=labels)
     return reducer.get_result()

pandas/core/apply.py

@@ -223,10 +223,12 @@ def apply_empty_result(self):
 
     def apply_raw(self):
         """ apply to the values as a numpy array """
-
         try:
             result = libreduction.compute_reduction(self.values, self.f, axis=self.axis)
-        except Exception:
+        except ValueError as err:
+            if "Function does not reduce" not in str(err):
+                # catch only ValueError raised intentionally in libreduction
+                raise
             result = np.apply_along_axis(self.f, self.axis, self.values)
 
         # TODO: mixed type case
@@ -273,24 +275,38 @@ def apply_standard(self):
         if (
             self.result_type in ["reduce", None]
             and not self.dtypes.apply(is_extension_type).any()
+            # Disallow complex_internals since libreduction shortcut
+            #  cannot handle MultiIndex
+            and not self.agg_axis._has_complex_internals
         ):
 
-            # Create a dummy Series from an empty array
-            from pandas import Series
-
             values = self.values
             index = self.obj._get_axis(self.axis)
             labels = self.agg_axis
             empty_arr = np.empty(len(index), dtype=values.dtype)
-            dummy = Series(empty_arr, index=index, dtype=values.dtype)
+
+            # Preserve subclass for e.g. test_subclassed_apply
+            dummy = self.obj._constructor_sliced(
+                empty_arr, index=index, dtype=values.dtype
+            )
 
             try:
                 result = libreduction.compute_reduction(
                     values, self.f, axis=self.axis, dummy=dummy, labels=labels
                 )
-                return self.obj._constructor_sliced(result, index=labels)
-            except Exception:
+            except ValueError as err:
+                if "Function does not reduce" not in str(err):
+                    # catch only ValueError raised intentionally in libreduction
+                    raise
+            except TypeError:
+                # e.g. test_apply_ignore_failures we just ignore
+                if not self.ignore_failures:
+                    raise
+            except ZeroDivisionError:
+                # reached via numexpr; fall back to python implementation
                 pass
+            else:
+                return self.obj._constructor_sliced(result, index=labels)
 
         # compute the result using the series generator
         self.apply_series_generator()

pandas/tests/groupby/test_groupby.py

@@ -775,11 +775,7 @@ def test_omit_nuisance(df):
 
     # won't work with axis = 1
     grouped = df.groupby({"A": 0, "C": 0, "D": 1, "E": 1}, axis=1)
-    msg = (
-        r'\("unsupported operand type\(s\) for \+: '
-        "'Timestamp' and 'float'\""
-        r", 'occurred at index 0'\)"
-    )
+    msg = r'\("unsupported operand type\(s\) for \+: ' "'Timestamp' and 'float'\", 0"
     with pytest.raises(TypeError, match=msg):
         grouped.agg(lambda x: x.sum(0, numeric_only=False))