BUG: FloatingArray * np.timedelta64 (#44772)

doc/source/whatsnew/v1.4.0.rst

@@ -644,6 +644,7 @@ Numeric
 - Bug in arithmetic operations involving :class:`RangeIndex` where the result would have the incorrect ``name`` (:issue:`43962`)
 - Bug in arithmetic operations involving :class:`Series` where the result could have the incorrect ``name`` when the operands having matching NA or matching tuple names (:issue:`44459`)
 - Bug in division with ``IntegerDtype`` or ``BooleanDtype`` array and NA scalar incorrectly raising (:issue:`44685`)
+- Bug in multiplying a :class:`Series` with ``FloatingDtype`` with a timedelta-like scalar incorrectly raising (:issue:`44772`)
 -
 
 Conversion

pandas/core/arrays/boolean.py

@@ -23,7 +23,6 @@
 
 from pandas.core.dtypes.common import (
     is_bool_dtype,
-    is_float,
     is_float_dtype,
     is_integer_dtype,
     is_list_like,
@@ -532,35 +531,5 @@ def _arith_method(self, other, op):
 
         return self._maybe_mask_result(result, mask, other, op_name)
 
-    def _maybe_mask_result(self, result, mask, other, op_name: str):
-        """
-        Parameters
-        ----------
-        result : array-like
-        mask : array-like bool
-        other : scalar or array-like
-        op_name : str
-        """
-        # if we have a float operand we are by-definition
-        # a float result
-        # or our op is a divide
-        if (is_float_dtype(other) or is_float(other)) or (
-            op_name in ["rtruediv", "truediv"]
-        ):
-            from pandas.core.arrays import FloatingArray
-
-            return FloatingArray(result, mask, copy=False)
-
-        elif is_bool_dtype(result):
-            return BooleanArray(result, mask, copy=False)
-
-        elif is_integer_dtype(result):
-            from pandas.core.arrays import IntegerArray
-
-            return IntegerArray(result, mask, copy=False)
-        else:
-            result[mask] = np.nan
-            return result
-
     def __abs__(self):
         return self.copy()

pandas/core/arrays/floating.py

@@ -354,27 +354,6 @@ def max(self, *, skipna=True, axis: int | None = 0, **kwargs):
         nv.validate_max((), kwargs)
         return super()._reduce("max", skipna=skipna, axis=axis)
 
-    def _maybe_mask_result(self, result, mask, other, op_name: str):
-        """
-        Parameters
-        ----------
-        result : array-like
-        mask : array-like bool
-        other : scalar or array-like
-        op_name : str
-        """
-        # TODO are there cases we don't end up with float?
-        # if we have a float operand we are by-definition
-        # a float result
-        # or our op is a divide
-        # if (is_float_dtype(other) or is_float(other)) or (
-        #     op_name in ["rtruediv", "truediv"]
-        # ):
-        #     result[mask] = np.nan
-        #     return result
-
-        return type(self)(result, mask, copy=False)
-
 
 _dtype_docstring = """
 An ExtensionDtype for {dtype} data.

pandas/core/arrays/integer.py

@@ -5,7 +5,6 @@
 import numpy as np
 
 from pandas._libs import (
-    iNaT,
     lib,
     missing as libmissing,
 )
@@ -26,7 +25,6 @@
 from pandas.core.dtypes.common import (
     is_bool_dtype,
     is_datetime64_dtype,
-    is_float,
     is_float_dtype,
     is_integer_dtype,
     is_object_dtype,
@@ -427,33 +425,6 @@ def max(self, *, skipna=True, axis: int | None = 0, **kwargs):
         nv.validate_max((), kwargs)
         return super()._reduce("max", skipna=skipna, axis=axis)
 
-    def _maybe_mask_result(self, result, mask, other, op_name: str):
-        """
-        Parameters
-        ----------
-        result : array-like
-        mask : array-like bool
-        other : scalar or array-like
-        op_name : str
-        """
-        # if we have a float operand we are by-definition
-        # a float result
-        # or our op is a divide
-        if (is_float_dtype(other) or is_float(other)) or (
-            op_name in ["rtruediv", "truediv"]
-        ):
-            from pandas.core.arrays import FloatingArray
-
-            return FloatingArray(result, mask, copy=False)
-
-        if result.dtype == "timedelta64[ns]":
-            from pandas.core.arrays import TimedeltaArray
-
-            result[mask] = iNaT
-            return TimedeltaArray._simple_new(result)
-
-        return type(self)(result, mask, copy=False)
-
 
 _dtype_docstring = """
 An ExtensionDtype for {dtype} integer data.

pandas/core/arrays/masked.py

@@ -12,6 +12,7 @@
 import numpy as np
 
 from pandas._libs import (
+    iNaT,
     lib,
     missing as libmissing,
 )
@@ -39,9 +40,11 @@
     is_bool,
     is_bool_dtype,
     is_dtype_equal,
+    is_float,
     is_float_dtype,
     is_integer_dtype,
     is_list_like,
+    is_numeric_dtype,
     is_object_dtype,
     is_scalar,
     is_string_dtype,
@@ -543,6 +546,48 @@ def _cmp_method(self, other, op) -> BooleanArray:
 
         return BooleanArray(result, mask, copy=False)
 
+    def _maybe_mask_result(self, result, mask, other, op_name: str):
+        """
+        Parameters
+        ----------
+        result : array-like
+        mask : array-like bool
+        other : scalar or array-like
+        op_name : str
+        """
+        # if we have a float operand we are by-definition
+        # a float result
+        # or our op is a divide
+        if (
+            (is_float_dtype(other) or is_float(other))
+            or (op_name in ["rtruediv", "truediv"])
+            or (is_float_dtype(self.dtype) and is_numeric_dtype(result.dtype))
+        ):
+            from pandas.core.arrays import FloatingArray
+
+            return FloatingArray(result, mask, copy=False)
+
+        elif is_bool_dtype(result):
+            from pandas.core.arrays import BooleanArray
+
+            return BooleanArray(result, mask, copy=False)
+
+        elif result.dtype == "timedelta64[ns]":
+            # e.g. test_numeric_arr_mul_tdscalar_numexpr_path
+            from pandas.core.arrays import TimedeltaArray
+
+            result[mask] = iNaT
+            return TimedeltaArray._simple_new(result)
+
+        elif is_integer_dtype(result):
+            from pandas.core.arrays import IntegerArray
+
+            return IntegerArray(result, mask, copy=False)
+
+        else:
+            result[mask] = np.nan
+            return result
+
     def isna(self) -> np.ndarray:
         return self._mask.copy()
 

pandas/core/arrays/numeric.py

@@ -14,7 +14,6 @@
     missing as libmissing,
 )
 from pandas.compat.numpy import function as nv
-from pandas.errors import AbstractMethodError
 
 from pandas.core.dtypes.common import (
     is_float,
@@ -80,9 +79,6 @@ class NumericArray(BaseMaskedArray):
     Base class for IntegerArray and FloatingArray.
     """
 
-    def _maybe_mask_result(self, result, mask, other, op_name: str):
-        raise AbstractMethodError(self)
-
     def _arith_method(self, other, op):
         op_name = op.__name__
         omask = None

pandas/tests/arithmetic/test_numeric.py

@@ -213,13 +213,18 @@ def test_numeric_arr_mul_tdscalar(self, scalar_td, numeric_idx, box_with_array):
         ],
         ids=lambda x: type(x).__name__,
     )
-    def test_numeric_arr_mul_tdscalar_numexpr_path(self, scalar_td, box_with_array):
+    @pytest.mark.parametrize("dtype", [np.int64, np.float64])
+    def test_numeric_arr_mul_tdscalar_numexpr_path(
+        self, dtype, scalar_td, box_with_array
+    ):
+        # GH#44772 for the float64 case
         box = box_with_array
 
-        arr = np.arange(2 * 10 ** 4).astype(np.int64)
+        arr_i8 = np.arange(2 * 10 ** 4).astype(np.int64, copy=False)
+        arr = arr_i8.astype(dtype, copy=False)
         obj = tm.box_expected(arr, box, transpose=False)
 
-        expected = arr.view("timedelta64[D]").astype("timedelta64[ns]")
+        expected = arr_i8.view("timedelta64[D]").astype("timedelta64[ns]")
         expected = tm.box_expected(expected, box, transpose=False)
 
         result = obj * scalar_td