CLN: cleanups in DataFrame._reduce (#36674)

Author: jbrockmendel

pandas/core/frame.py

@@ -638,10 +638,10 @@ def _can_fast_transpose(self) -> bool:
         """
         Can we transpose this DataFrame without creating any new array objects.
         """
-        if self._data.any_extension_types:
+        if self._mgr.any_extension_types:
             # TODO(EA2D) special case would be unnecessary with 2D EAs
             return False
-        return len(self._data.blocks) == 1
+        return len(self._mgr.blocks) == 1
 
     # ----------------------------------------------------------------------
     # Rendering Methods
@@ -2879,7 +2879,7 @@ def _get_column_array(self, i: int) -> ArrayLike:
         Get the values of the i'th column (ndarray or ExtensionArray, as stored
         in the Block)
         """
-        return self._data.iget_values(i)
+        return self._mgr.iget_values(i)
 
     def _iter_column_arrays(self) -> Iterator[ArrayLike]:
         """
@@ -4911,7 +4911,7 @@ def _maybe_casted_values(index, labels=None):
 
     @doc(NDFrame.isna, klass=_shared_doc_kwargs["klass"])
     def isna(self) -> DataFrame:
-        result = self._constructor(self._data.isna(func=isna))
+        result = self._constructor(self._mgr.isna(func=isna))
         return result.__finalize__(self, method="isna")
 
     @doc(NDFrame.isna, klass=_shared_doc_kwargs["klass"])
@@ -8575,6 +8575,7 @@ def _reduce(
     ):
 
         assert filter_type is None or filter_type == "bool", filter_type
+        out_dtype = "bool" if filter_type == "bool" else None
 
         dtype_is_dt = np.array(
             [
@@ -8594,10 +8595,9 @@ def _reduce(
             cols = self.columns[~dtype_is_dt]
             self = self[cols]
 
-        # TODO: Make other agg func handle axis=None properly
+        # TODO: Make other agg func handle axis=None properly GH#21597
         axis = self._get_axis_number(axis)
         labels = self._get_agg_axis(axis)
-        constructor = self._constructor
         assert axis in [0, 1]
 
         def func(values):
@@ -8606,18 +8606,19 @@ def func(values):
             else:
                 return op(values, axis=axis, skipna=skipna, **kwds)
 
+        def blk_func(values):
+            if isinstance(values, ExtensionArray):
+                return values._reduce(name, skipna=skipna, **kwds)
+            else:
+                return op(values, axis=1, skipna=skipna, **kwds)
+
         def _get_data() -> DataFrame:
             if filter_type is None:
                 data = self._get_numeric_data()
-            elif filter_type == "bool":
+            else:
                 # GH#25101, GH#24434
+                assert filter_type == "bool"
                 data = self._get_bool_data()
-            else:  # pragma: no cover
-                msg = (
-                    f"Generating numeric_only data with filter_type {filter_type} "
-                    "not supported."
-                )
-                raise NotImplementedError(msg)
             return data
 
         if numeric_only is not None:
@@ -8628,14 +8629,6 @@ def _get_data() -> DataFrame:
                 df = df.T
                 axis = 0
 
-            out_dtype = "bool" if filter_type == "bool" else None
-
-            def blk_func(values):
-                if isinstance(values, ExtensionArray):
-                    return values._reduce(name, skipna=skipna, **kwds)
-                else:
-                    return op(values, axis=1, skipna=skipna, **kwds)
-
             # After possibly _get_data and transposing, we are now in the
             #  simple case where we can use BlockManager.reduce
             res = df._mgr.reduce(blk_func)
@@ -8651,11 +8644,10 @@ def blk_func(values):
         if not self._is_homogeneous_type or self._mgr.any_extension_types:
             # try to avoid self.values call
 
-            if filter_type is None and axis == 0 and len(self) > 0:
+            if filter_type is None and axis == 0:
                 # operate column-wise
 
                 # numeric_only must be None here, as other cases caught above
-                # require len(self) > 0 bc frame_apply messes up empty prod/sum
 
                 # this can end up with a non-reduction
                 # but not always. if the types are mixed
@@ -8691,19 +8683,17 @@ def blk_func(values):
             with np.errstate(all="ignore"):
                 result = func(values)
 
-        if is_object_dtype(result.dtype):
+        if filter_type == "bool" and notna(result).all():
+            result = result.astype(np.bool_)
+        elif filter_type is None and is_object_dtype(result.dtype):
             try:
-                if filter_type is None:
-                    result = result.astype(np.float64)
-                elif filter_type == "bool" and notna(result).all():
-                    result = result.astype(np.bool_)
+                result = result.astype(np.float64)
             except (ValueError, TypeError):
                 # try to coerce to the original dtypes item by item if we can
                 if axis == 0:
                     result = coerce_to_dtypes(result, data.dtypes)
 
-        if constructor is not None:
-            result = self._constructor_sliced(result, index=labels)
+        result = self._constructor_sliced(result, index=labels)
         return result
 
     def nunique(self, axis=0, dropna=True) -> Series: