BUG: inconsistent DataFrame.agg behavoir when passing as kwargs numeric_only=True

doc/source/whatsnew/v1.5.3.rst

@@ -24,6 +24,7 @@ Fixed regressions
 
 Bug fixes
 ~~~~~~~~~
+- Bug in inconsistent DataFrame.agg behavior when passing as kwargs ``numeric_only=True`` with their associated string alias as functions (:issue:`49352`)
 - Bug in :meth:`.Styler.to_excel` leading to error when unrecognized ``border-style`` (e.g. ``"hair"``) provided to Excel writers (:issue:`48649`)
 -
 

pandas/core/apply.py

@@ -64,7 +64,6 @@
     from pandas.core.resample import Resampler
     from pandas.core.window.rolling import BaseWindow
 
-
 ResType = Dict[int, Any]
 
 
@@ -284,6 +283,12 @@ def transform_str_or_callable(self, func) -> DataFrame | Series:
         except Exception:
             return func(obj, *args, **kwargs)
 
+    def _maybe_filter_numeric_only(self) -> Any:
+        if "numeric_only" in self.kwargs and self.kwargs["numeric_only"] is True:
+            obj = self.obj._get_numeric_data()
+            return obj
+        return self.obj
+
     def agg_list_like(self) -> DataFrame | Series:
         """
         Compute aggregation in the case of a list-like argument.
@@ -294,7 +299,7 @@ def agg_list_like(self) -> DataFrame | Series:
         """
         from pandas.core.reshape.concat import concat
 
-        obj = self.obj
+        obj = self._maybe_filter_numeric_only()
         arg = cast(List[AggFuncTypeBase], self.f)
 
         if getattr(obj, "axis", 0) == 1:
@@ -1169,7 +1174,6 @@ def reconstruct_func(
 
     if not relabeling:
         if isinstance(func, list) and len(func) > len(set(func)):
-
             # GH 28426 will raise error if duplicated function names are used and
             # there is no reassigned name
             raise SpecificationError(

pandas/core/frame.py

@@ -9213,6 +9213,21 @@ def _gotitem(
     2    8.0
     3    NaN
     dtype: float64
+
+    Aggregate functions on DataFrames with mixed non-numeric and numeric columns.
+
+    >>> df = pd.DataFrame([['a', 1, 4],
+    ...                    ['b', 2, 5],
+    ...                    ['c', 3, 6]],
+    ...                   columns=['A', 'B', 'C'])
+
+    Works equivalently as above. Add argument `numeric_only=True` to
+    aggregate only numeric columns.
+
+    >>> df.agg(['mean', 'std'], numeric_only=True)
+            B     C
+    mean  2.0   5.0
+    std   1.0   1.0
     """
     )
 
@@ -9377,6 +9392,9 @@ def apply(
         Functions that mutate the passed object can produce unexpected
         behavior or errors and are not supported. See :ref:`gotchas.udf-mutation`
         for more details.
+        Use the keyword argument `numeric_only=True` to apply functions
+        only to numeric columns and to skip the non-numeric columns,
+        e.g. the column contains a string.
 
         Examples
         --------

pandas/core/shared_docs.py

@@ -46,6 +46,10 @@
 for more details.
 
 A passed user-defined-function will be passed a Series for evaluation.
+
+Use the keyword argument `numeric_only=True` to apply functions
+only to numeric columns and to skip the non-numeric columns,
+e.g. the column contains a string.
 {examples}"""
 
 _shared_docs[

pandas/tests/apply/test_frame_apply.py

@@ -858,6 +858,22 @@ def test_with_dictlike_columns_with_infer():
     tm.assert_frame_equal(result, expected)
 
 
+def test_with_dictlike_functions():
+    # GH 49352
+
+    df = DataFrame({"a": [1, 2, 3], "b": list("abc"), "c": [3, 4, 5]})
+    functions = {"a": "std", "c": ["mean", "sum"]}
+    result = df.agg(functions, numeric_only=True)
+    expected = DataFrame(
+        index=["std", "mean", "sum"],
+        data={
+            "a": [1.0, np.nan, np.nan],
+            "c": [np.nan, 4.0, 12.0],
+        },
+    )
+    tm.assert_frame_equal(result, expected)
+
+
 def test_with_listlike_columns():
     # GH 17348
     df = DataFrame(
@@ -887,6 +903,24 @@ def test_with_listlike_columns_returning_list():
     tm.assert_series_equal(result, expected)
 
 
+def test_with_listlike_functions():
+    # GH 49352
+
+    df = DataFrame({"a": [1, 2, 3], "b": list("abc"), "c": [3, 4, 5]})
+    functions = ["mean", "std"]
+    result = df.agg(functions, numeric_only=True)
+    expected = DataFrame(index=["mean", "std"], data={"a": [2.0, 1.0], "c": [4.0, 1.0]})
+    tm.assert_frame_equal(result, expected)
+
+    expected = df.agg("mean", numeric_only=True)
+    expected.name = "mean"
+    tm.assert_series_equal(result.loc["mean"], expected)
+
+    expected = df.agg("std", numeric_only=True)
+    expected.name = "std"
+    tm.assert_series_equal(result.loc["std"], expected)
+
+
 def test_infer_output_shape_columns():
     # GH 18573