DEPR: Enforce deprecation of numeric_only=None in DataFrame aggregations (pandas-dev#49551)

* WIP

* DEPR: Enforce deprecation of numeric_only=None in DataFrame aggregations

* Partial reverts

* numeric_only in generic/series, fixup

* cleanup

* Remove docs warning

* fixups

* Fixups

Author: rhshadrach

doc/source/whatsnew/v1.2.0.rst

@@ -383,12 +383,17 @@ this pathological behavior (:issue:`37827`):
 
 *New behavior*:
 
-.. ipython:: python
-   :okwarning:
+.. code-block:: ipython
 
-    df.mean()
+    In [3]: df.mean()
+    Out[3]:
+    A    1.0
+    dtype: float64
 
-    df[["A"]].mean()
+    In [4]: df[["A"]].mean()
+    Out[4]:
+    A    1.0
+    dtype: float64
 
 Moreover, DataFrame reductions with ``numeric_only=None`` will now be
 consistent with their Series counterparts.  In particular, for
@@ -415,10 +420,10 @@ instead of casting to a NumPy array which may have different semantics (:issue:`
 
 *New behavior*:
 
-.. ipython:: python
-   :okwarning:
+.. code-block:: ipython
 
-    df.any()
+    In [5]: df.any()
+    Out[5]: Series([], dtype: bool)
 
 
 .. _whatsnew_120.api_breaking.python:

doc/source/whatsnew/v2.0.0.rst

@@ -501,6 +501,7 @@ Removal of prior version deprecations/changes
 - Changed behavior of :meth:`Series.__setitem__` with an integer key and a :class:`Float64Index` when the key is not present in the index; previously we treated the key as positional (behaving like ``series.iloc[key] = val``), now we treat it is a label (behaving like ``series.loc[key] = val``), consistent with :meth:`Series.__getitem__`` behavior (:issue:`33469`)
 - Removed ``na_sentinel`` argument from :func:`factorize`, :meth:`.Index.factorize`, and :meth:`.ExtensionArray.factorize` (:issue:`47157`)
 - Changed behavior of :meth:`DataFrameGroupBy.apply` and :meth:`SeriesGroupBy.apply` so that ``group_keys`` is respected even if a transformer is detected (:issue:`34998`)
+- Enforced deprecation ``numeric_only=None`` (the default) in DataFrame reductions that would silently drop columns that raised; ``numeric_only`` now defaults to ``False`` (:issue:`41480`)
 -
 
 .. ---------------------------------------------------------------------------
@@ -570,6 +571,7 @@ Timezones
 Numeric
 ^^^^^^^
 - Bug in :meth:`DataFrame.add` cannot apply ufunc when inputs contain mixed DataFrame type and Series type (:issue:`39853`)
+- Bug in DataFrame reduction methods (e.g. :meth:`DataFrame.sum`) with object dtype, ``axis=1`` and ``numeric_only=False`` would not be coerced to float (:issue:`49551`)
 -
 
 Conversion

pandas/core/frame.py

@@ -266,9 +266,8 @@
     you to specify a location to update with some value.""",
 }
 
-_numeric_only_doc = """numeric_only : bool or None, default None
-    Include only float, int, boolean data. If None, will attempt to use
-    everything, then use only numeric data
+_numeric_only_doc = """numeric_only : bool, default False
+    Include only float, int, boolean data.
 """
 
 _merge_doc = """
@@ -10489,7 +10488,7 @@ def _reduce(
         *,
         axis: Axis = 0,
         skipna: bool = True,
-        numeric_only: bool | None = None,
+        numeric_only: bool = False,
         filter_type=None,
         **kwds,
     ):
@@ -10498,7 +10497,6 @@ def _reduce(
 
         # TODO: Make other agg func handle axis=None properly GH#21597
         axis = self._get_axis_number(axis)
-        labels = self._get_agg_axis(axis)
         assert axis in [0, 1]
 
         def func(values: np.ndarray):
@@ -10524,25 +10522,22 @@ def _get_data() -> DataFrame:
                 data = self._get_bool_data()
             return data
 
-        numeric_only_bool = com.resolve_numeric_only(numeric_only)
-        if numeric_only is not None or axis == 0:
+        if numeric_only or axis == 0:
             # For numeric_only non-None and axis non-None, we know
             #  which blocks to use and no try/except is needed.
             #  For numeric_only=None only the case with axis==0 and no object
             #  dtypes are unambiguous can be handled with BlockManager.reduce
             # Case with EAs see GH#35881
             df = self
-            if numeric_only_bool:
+            if numeric_only:
                 df = _get_data()
             if axis == 1:
                 df = df.T
                 axis = 0
 
-            ignore_failures = numeric_only is None
-
             # 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, ignore_failures=ignore_failures)
+            res, _ = df._mgr.reduce(blk_func, ignore_failures=False)
             out = df._constructor(res).iloc[0]
             if out_dtype is not None:
                 out = out.astype(out_dtype)
@@ -10559,36 +10554,11 @@ def _get_data() -> DataFrame:
 
             return out
 
-        assert numeric_only is None
+        assert not numeric_only and axis == 1
 
         data = self
         values = data.values
-
-        try:
-            result = func(values)
-
-        except TypeError:
-            # e.g. in nanops trying to convert strs to float
-
-            data = _get_data()
-            labels = data._get_agg_axis(axis)
-
-            values = data.values
-            with np.errstate(all="ignore"):
-                result = func(values)
-
-            # columns have been dropped GH#41480
-            arg_name = "numeric_only"
-            if name in ["all", "any"]:
-                arg_name = "bool_only"
-            warnings.warn(
-                "Dropping of nuisance columns in DataFrame reductions "
-                f"(with '{arg_name}=None') is deprecated; in a future "
-                "version this will raise TypeError.  Select only valid "
-                "columns before calling the reduction.",
-                FutureWarning,
-                stacklevel=find_stack_level(),
-            )
+        result = func(values)
 
         if hasattr(result, "dtype"):
             if filter_type == "bool" and notna(result).all():
@@ -10600,6 +10570,7 @@ def _get_data() -> DataFrame:
                     # try to coerce to the original dtypes item by item if we can
                     pass
 
+        labels = self._get_agg_axis(axis)
         result = self._constructor_sliced(result, index=labels)
         return result