REF: de-duplicate ndarray[datetimelike] wrapping

pandas/core/arrays/interval.py

@@ -44,7 +44,7 @@
 from pandas.core.arrays.base import ExtensionArray, _extension_array_shared_docs
 from pandas.core.arrays.categorical import Categorical
 import pandas.core.common as com
-from pandas.core.construction import array, extract_array
+from pandas.core.construction import array, extract_array, wrap_if_datetimelike
 from pandas.core.indexers import check_array_indexer
 from pandas.core.indexes.base import ensure_index
 from pandas.core.ops import invalid_comparison, unpack_zerodim_and_defer
@@ -251,11 +251,9 @@ def _simple_new(
             raise ValueError(msg)
 
         # For dt64/td64 we want DatetimeArray/TimedeltaArray instead of ndarray
-        from pandas.core.ops.array_ops import maybe_upcast_datetimelike_array
-
-        left = maybe_upcast_datetimelike_array(left)
+        left = wrap_if_datetimelike(left)
         left = extract_array(left, extract_numpy=True)
-        right = maybe_upcast_datetimelike_array(right)
+        right = wrap_if_datetimelike(right)
         right = extract_array(right, extract_numpy=True)
 
         lbase = getattr(left, "_ndarray", left).base

pandas/core/construction.py

@@ -402,6 +402,24 @@ def extract_array(obj: object, extract_numpy: bool = False) -> Union[Any, ArrayL
     return obj
 
 
+def wrap_if_datetimelike(arr):
+    """
+    Wrap datetime64 and timedelta64 ndarrays in DatetimeArray/TimedeltaArray.
+    """
+    if isinstance(arr, np.ndarray):
+        if arr.dtype.kind == "M":
+            from pandas.core.arrays import DatetimeArray
+
+            return DatetimeArray._from_sequence(arr)
+
+        elif arr.dtype.kind == "m":
+            from pandas.core.arrays import TimedeltaArray
+
+            return TimedeltaArray._from_sequence(arr)
+
+    return arr
+
+
 def sanitize_array(
     data,
     index: Optional[Index],

pandas/core/dtypes/concat.py

@@ -18,7 +18,7 @@
 
 from pandas.core.arrays import ExtensionArray
 from pandas.core.arrays.sparse import SparseArray
-from pandas.core.construction import array
+from pandas.core.construction import array, wrap_if_datetimelike
 
 
 def _get_dtype_kinds(arrays) -> Set[str]:
@@ -364,12 +364,13 @@ def _concat_datetime(to_concat, axis=0, typs=None):
     if typs is None:
         typs = _get_dtype_kinds(to_concat)
 
-    to_concat = [_wrap_datetimelike(x) for x in to_concat]
+    to_concat = [wrap_if_datetimelike(x) for x in to_concat]
     single_dtype = len({x.dtype for x in to_concat}) == 1
 
     # multiple types, need to coerce to object
     if not single_dtype:
-        # wrap_datetimelike ensures that astype(object) wraps in Timestamp/Timedelta
+        # wrap_if_datetimelike ensures that astype(object) wraps
+        #  in Timestamp/Timedelta
         return _concatenate_2d([x.astype(object) for x in to_concat], axis=axis)
 
     if axis == 1:
@@ -383,17 +384,3 @@ def _concat_datetime(to_concat, axis=0, typs=None):
         assert result.shape[0] == 1
         result = result[0]
     return result
-
-
-def _wrap_datetimelike(arr):
-    """
-    Wrap datetime64 and timedelta64 ndarrays in DatetimeArray/TimedeltaArray.
-
-    DTA/TDA handle .astype(object) correctly.
-    """
-    from pandas.core.construction import array as pd_array, extract_array
-
-    arr = extract_array(arr, extract_numpy=True)
-    if isinstance(arr, np.ndarray) and arr.dtype.kind in ["m", "M"]:
-        arr = pd_array(arr)
-    return arr

pandas/core/ops/array_ops.py

@@ -30,6 +30,7 @@
 from pandas.core.dtypes.generic import ABCExtensionArray, ABCIndexClass, ABCSeries
 from pandas.core.dtypes.missing import isna, notna
 
+from pandas.core.construction import wrap_if_datetimelike
 from pandas.core.ops import missing
 from pandas.core.ops.dispatch import should_extension_dispatch
 from pandas.core.ops.invalid import invalid_comparison
@@ -175,8 +176,8 @@ def arithmetic_op(left: ArrayLike, right: Any, op):
 
     # NB: We assume that extract_array has already been called
     #  on `left` and `right`.
-    lvalues = maybe_upcast_datetimelike_array(left)
-    rvalues = maybe_upcast_datetimelike_array(right)
+    lvalues = wrap_if_datetimelike(left)
+    rvalues = wrap_if_datetimelike(right)
     rvalues = _maybe_upcast_for_op(rvalues, lvalues.shape)
 
     if should_extension_dispatch(lvalues, rvalues) or isinstance(rvalues, Timedelta):
@@ -206,7 +207,7 @@ def comparison_op(left: ArrayLike, right: Any, op) -> ArrayLike:
     ndarray or ExtensionArray
     """
     # NB: We assume extract_array has already been called on left and right
-    lvalues = maybe_upcast_datetimelike_array(left)
+    lvalues = wrap_if_datetimelike(left)
     rvalues = right
 
     rvalues = lib.item_from_zerodim(rvalues)
@@ -331,7 +332,7 @@ def fill_bool(x, left=None):
         right = construct_1d_object_array_from_listlike(right)
 
     # NB: We assume extract_array has already been called on left and right
-    lvalues = maybe_upcast_datetimelike_array(left)
+    lvalues = wrap_if_datetimelike(left)
     rvalues = right
 
     if should_extension_dispatch(lvalues, rvalues):
@@ -400,31 +401,6 @@ def get_array_op(op):
         raise NotImplementedError(op_name)
 
 
-def maybe_upcast_datetimelike_array(obj: ArrayLike) -> ArrayLike:
-    """
-    If we have an ndarray that is either datetime64 or timedelta64, wrap in EA.
-
-    Parameters
-    ----------
-    obj : ndarray or ExtensionArray
-
-    Returns
-    -------
-    ndarray or ExtensionArray
-    """
-    if isinstance(obj, np.ndarray):
-        if obj.dtype.kind == "m":
-            from pandas.core.arrays import TimedeltaArray
-
-            return TimedeltaArray._from_sequence(obj)
-        if obj.dtype.kind == "M":
-            from pandas.core.arrays import DatetimeArray
-
-            return DatetimeArray._from_sequence(obj)
-
-    return obj
-
-
 def _maybe_upcast_for_op(obj, shape: Shape):
     """
     Cast non-pandas objects to pandas types to unify behavior of arithmetic