DEPR: support axis=None in DataFrame reductions (pandas-dev#52042)

* DEPR: support axis=None in DataFrame reductions

* test, whatsnew

* catch in apply(sum)

* Fix defaults

* catch warnings

* dont check stacklevel

* mypy fixup

* catch warning

Author: jbrockmendel

doc/source/whatsnew/v0.15.1.rst

@@ -102,6 +102,7 @@ API changes
   current behavior:
 
   .. ipython:: python
+     :okwarning:
 
      gr.apply(sum)
 

doc/source/whatsnew/v2.1.0.rst

@@ -253,6 +253,7 @@ Deprecations
 - Deprecated :meth:`DataFrame.applymap`. Use the new :meth:`DataFrame.map` method instead (:issue:`52353`)
 - Deprecated :meth:`DataFrame.swapaxes` and :meth:`Series.swapaxes`, use :meth:`DataFrame.transpose` or :meth:`Series.transpose` instead (:issue:`51946`)
 - Deprecated ``freq`` parameter in :class:`PeriodArray` constructor, pass ``dtype`` instead (:issue:`52462`)
+- Deprecated behavior of :class:`DataFrame` reductions ``sum``, ``prod``, ``std``, ``var``, ``sem`` with ``axis=None``, in a future version this will operate over both axes returning a scalar instead of behaving like ``axis=0``; note this also affects numpy functions e.g. ``np.sum(df)`` (:issue:`21597`)
 - Deprecated behavior of :func:`concat` when :class:`DataFrame` has columns that are all-NA, in a future version these will not be discarded when determining the resulting dtype (:issue:`40893`)
 - Deprecated behavior of :meth:`Series.dt.to_pydatetime`, in a future version this will return a :class:`Series` containing python ``datetime`` objects instead of an ``ndarray`` of datetimes; this matches the behavior of other :meth:`Series.dt` properties (:issue:`20306`)
 - Deprecated logical operations (``|``, ``&``, ``^``) between pandas objects and dtype-less sequences (e.g. ``list``, ``tuple``), wrap a sequence in a :class:`Series` or numpy array before operating instead (:issue:`51521`)

pandas/core/frame.py

@@ -10981,7 +10981,7 @@ def max(
     @doc(make_doc("sum", ndim=2))
     def sum(
         self,
-        axis: Axis | None = None,
+        axis: Axis | None = 0,
         skipna: bool = True,
         numeric_only: bool = False,
         min_count: int = 0,
@@ -10993,7 +10993,7 @@ def sum(
     @doc(make_doc("prod", ndim=2))
     def prod(
         self,
-        axis: Axis | None = None,
+        axis: Axis | None = 0,
         skipna: bool = True,
         numeric_only: bool = False,
         min_count: int = 0,
@@ -11024,7 +11024,7 @@ def median(
     @doc(make_doc("sem", ndim=2))
     def sem(
         self,
-        axis: Axis | None = None,
+        axis: Axis | None = 0,
         skipna: bool = True,
         ddof: int = 1,
         numeric_only: bool = False,
@@ -11035,7 +11035,7 @@ def sem(
     @doc(make_doc("var", ndim=2))
     def var(
         self,
-        axis: Axis | None = None,
+        axis: Axis | None = 0,
         skipna: bool = True,
         ddof: int = 1,
         numeric_only: bool = False,
@@ -11046,7 +11046,7 @@ def var(
     @doc(make_doc("std", ndim=2))
     def std(
         self,
-        axis: Axis | None = None,
+        axis: Axis | None = 0,
         skipna: bool = True,
         ddof: int = 1,
         numeric_only: bool = False,

pandas/core/generic.py

@@ -11290,6 +11290,8 @@ def _logical_func(
                 name, func, axis=0, bool_only=bool_only, skipna=skipna, **kwargs
             )
             return res._logical_func(name, func, skipna=skipna, **kwargs)
+        elif axis is None:
+            axis = 0
 
         if (
             self.ndim > 1
@@ -11394,15 +11396,27 @@ def _stat_function_ddof(
         self,
         name: str,
         func,
-        axis: Axis | None = None,
+        axis: Axis | None | lib.NoDefault = lib.no_default,
         skipna: bool_t = True,
         ddof: int = 1,
         numeric_only: bool_t = False,
         **kwargs,
     ) -> Series | float:
         nv.validate_stat_ddof_func((), kwargs, fname=name)
         validate_bool_kwarg(skipna, "skipna", none_allowed=False)
+
         if axis is None:
+            if self.ndim > 1:
+                warnings.warn(
+                    f"The behavior of {type(self).__name__}.{name} with axis=None "
+                    "is deprecated, in a future version this will reduce over both "
+                    "axes and return a scalar. To retain the old behavior, pass "
+                    "axis=0 (or do not pass axis)",
+                    FutureWarning,
+                    stacklevel=find_stack_level(),
+                )
+            axis = 0
+        elif axis is lib.no_default:
             axis = 0
 
         return self._reduce(
@@ -11411,7 +11425,7 @@ def _stat_function_ddof(
 
     def sem(
         self,
-        axis: Axis | None = None,
+        axis: Axis | None = 0,
         skipna: bool_t = True,
         ddof: int = 1,
         numeric_only: bool_t = False,
@@ -11423,7 +11437,7 @@ def sem(
 
     def var(
         self,
-        axis: Axis | None = None,
+        axis: Axis | None = 0,
         skipna: bool_t = True,
         ddof: int = 1,
         numeric_only: bool_t = False,
@@ -11435,7 +11449,7 @@ def var(
 
     def std(
         self,
-        axis: Axis | None = None,
+        axis: Axis | None = 0,
         skipna: bool_t = True,
         ddof: int = 1,
         numeric_only: bool_t = False,
@@ -11547,7 +11561,7 @@ def _min_count_stat_function(
         self,
         name: str,
         func,
-        axis: Axis | None = None,
+        axis: Axis | None | lib.NoDefault = lib.no_default,
         skipna: bool_t = True,
         numeric_only: bool_t = False,
         min_count: int = 0,
@@ -11559,6 +11573,17 @@ def _min_count_stat_function(
         validate_bool_kwarg(skipna, "skipna", none_allowed=False)
 
         if axis is None:
+            if self.ndim > 1:
+                warnings.warn(
+                    f"The behavior of {type(self).__name__}.{name} with axis=None "
+                    "is deprecated, in a future version this will reduce over both "
+                    "axes and return a scalar. To retain the old behavior, pass "
+                    "axis=0 (or do not pass axis)",
+                    FutureWarning,
+                    stacklevel=find_stack_level(),
+                )
+            axis = 0
+        elif axis is lib.no_default:
             axis = 0
 
         return self._reduce(
@@ -11572,7 +11597,7 @@ def _min_count_stat_function(
 
     def sum(
         self,
-        axis: Axis | None = None,
+        axis: Axis | None = 0,
         skipna: bool_t = True,
         numeric_only: bool_t = False,
         min_count: int = 0,
@@ -11584,7 +11609,7 @@ def sum(
 
     def prod(
         self,
-        axis: Axis | None = None,
+        axis: Axis | None = 0,
         skipna: bool_t = True,
         numeric_only: bool_t = False,
         min_count: int = 0,

pandas/tests/frame/test_npfuncs.py

@@ -27,6 +27,23 @@ def test_np_sqrt(self, float_frame):
 
         tm.assert_frame_equal(result, float_frame.apply(np.sqrt))
 
+    def test_sum_deprecated_axis_behavior(self):
+        # GH#52042 deprecated behavior of df.sum(axis=None), which gets
+        #  called when we do np.sum(df)
+
+        arr = np.random.randn(4, 3)
+        df = DataFrame(arr)
+
+        msg = "The behavior of DataFrame.sum with axis=None is deprecated"
+        with tm.assert_produces_warning(
+            FutureWarning, match=msg, check_stacklevel=False
+        ):
+            res = np.sum(df)
+
+        with tm.assert_produces_warning(FutureWarning, match=msg):
+            expected = df.sum(axis=None)
+        tm.assert_series_equal(res, expected)
+
     def test_np_ravel(self):
         # GH26247
         arr = np.array(

pandas/tests/groupby/aggregate/test_aggregate.py

@@ -153,7 +153,11 @@ def test_agg_apply_corner(ts, tsframe):
     )
     tm.assert_frame_equal(grouped.sum(), exp_df)
     tm.assert_frame_equal(grouped.agg(np.sum), exp_df)
-    tm.assert_frame_equal(grouped.apply(np.sum), exp_df)
+
+    msg = "The behavior of DataFrame.sum with axis=None is deprecated"
+    with tm.assert_produces_warning(FutureWarning, match=msg, check_stacklevel=False):
+        res = grouped.apply(np.sum)
+    tm.assert_frame_equal(res, exp_df)
 
 
 def test_agg_grouping_is_list_tuple(ts):

pandas/tests/groupby/test_apply.py

@@ -1072,14 +1072,17 @@ def test_apply_is_unchanged_when_other_methods_are_called_first(reduction_func):
 
     # Check output when no other methods are called before .apply()
     grp = df.groupby(by="a")
-    result = grp.apply(sum)
+    msg = "The behavior of DataFrame.sum with axis=None is deprecated"
+    with tm.assert_produces_warning(FutureWarning, match=msg, check_stacklevel=False):
+        result = grp.apply(sum)
     tm.assert_frame_equal(result, expected)
 
     # Check output when another method is called before .apply()
     grp = df.groupby(by="a")
     args = get_groupby_method_args(reduction_func, df)
     _ = getattr(grp, reduction_func)(*args)
-    result = grp.apply(sum)
+    with tm.assert_produces_warning(FutureWarning, match=msg, check_stacklevel=False):
+        result = grp.apply(sum)
     tm.assert_frame_equal(result, expected)
 
 

pandas/tests/groupby/test_function.py

@@ -73,7 +73,11 @@ def test_builtins_apply(keys, f):
     gb = df.groupby(keys)
 
     fname = f.__name__
-    result = gb.apply(f)
+
+    warn = None if f is not sum else FutureWarning
+    msg = "The behavior of DataFrame.sum with axis=None is deprecated"
+    with tm.assert_produces_warning(warn, match=msg, check_stacklevel=False):
+        result = gb.apply(f)
     ngroups = len(df.drop_duplicates(subset=keys))
 
     assert_msg = f"invalid frame shape: {result.shape} (expected ({ngroups}, 3))"

pandas/tests/groupby/test_groupby.py

@@ -752,7 +752,13 @@ def test_groupby_as_index_agg(df):
 
     gr = df.groupby(ts)
     gr.nth(0)  # invokes set_selection_from_grouper internally
-    tm.assert_frame_equal(gr.apply(sum), df.groupby(ts).apply(sum))
+
+    msg = "The behavior of DataFrame.sum with axis=None is deprecated"
+    with tm.assert_produces_warning(FutureWarning, match=msg, check_stacklevel=False):
+        res = gr.apply(sum)
+    with tm.assert_produces_warning(FutureWarning, match=msg, check_stacklevel=False):
+        alt = df.groupby(ts).apply(sum)
+    tm.assert_frame_equal(res, alt)
 
     for attr in ["mean", "max", "count", "idxmax", "cumsum", "all"]:
         gr = df.groupby(ts, as_index=False)

pandas/tests/window/test_expanding.py

@@ -333,7 +333,12 @@ def test_expanding_func(func, static_comp, frame_or_series):
     result = getattr(obj, func)()
     assert isinstance(result, frame_or_series)
 
-    expected = static_comp(data[:11])
+    msg = "The behavior of DataFrame.sum with axis=None is deprecated"
+    warn = None
+    if frame_or_series is DataFrame and static_comp is np.sum:
+        warn = FutureWarning
+    with tm.assert_produces_warning(warn, match=msg, check_stacklevel=False):
+        expected = static_comp(data[:11])
     if frame_or_series is Series:
         tm.assert_almost_equal(result[10], expected)
     else: