BUG: Index.view with non-nano

doc/source/whatsnew/v2.2.0.rst

@@ -321,6 +321,7 @@ Datetimelike
 ^^^^^^^^^^^^
 - Bug in :func:`concat` raising ``AttributeError`` when concatenating all-NA DataFrame with :class:`DatetimeTZDtype` dtype DataFrame. (:issue:`52093`)
 - Bug in :meth:`DatetimeIndex.union` returning object dtype for tz-aware indexes with the same timezone but different units (:issue:`55238`)
+- Bug in :meth:`Index.view` to a datetime64 dtype with non-supported resolution incorrectly raising (:issue:`55710`)
 - Bug in :meth:`Tick.delta` with very large ticks raising ``OverflowError`` instead of ``OutOfBoundsTimedelta`` (:issue:`55503`)
 - Bug in adding or subtracting a :class:`Week` offset to a ``datetime64`` :class:`Series`, :class:`Index`, or :class:`DataFrame` column with non-nanosecond resolution returning incorrect results (:issue:`55583`)
 - Bug in addition or subtraction of :class:`BusinessDay` offset with ``offset`` attribute to non-nanosecond :class:`Index`, :class:`Series`, or :class:`DataFrame` column giving incorrect results (:issue:`55608`)

pandas/core/arrays/_mixins.py

@@ -13,6 +13,10 @@
 
 from pandas._libs import lib
 from pandas._libs.arrays import NDArrayBacked
+from pandas._libs.tslibs import (
+    get_unit_from_dtype,
+    is_supported_unit,
+)
 from pandas._typing import (
     ArrayLike,
     AxisInt,
@@ -137,21 +141,19 @@ def view(self, dtype: Dtype | None = None) -> ArrayLike:
             cls = dtype.construct_array_type()  # type: ignore[assignment]
             dt64_values = arr.view(f"M8[{dtype.unit}]")
             return cls(dt64_values, dtype=dtype)
-        elif dtype == "M8[ns]":
+        elif lib.is_np_dtype(dtype, "M") and is_supported_unit(
+            get_unit_from_dtype(dtype)
+        ):
             from pandas.core.arrays import DatetimeArray
 
-            # error: Argument 1 to "view" of "ndarray" has incompatible type
-            # "ExtensionDtype | dtype[Any]"; expected "dtype[Any] | type[Any]
-            # | _SupportsDType[dtype[Any]]"
-            dt64_values = arr.view(dtype)  # type: ignore[arg-type]
+            dt64_values = arr.view(dtype)
             return DatetimeArray(dt64_values, dtype=dtype)
-        elif dtype == "m8[ns]":
+        elif lib.is_np_dtype(dtype, "m") and is_supported_unit(
+            get_unit_from_dtype(dtype)
+        ):
             from pandas.core.arrays import TimedeltaArray
 
-            # error: Argument 1 to "view" of "ndarray" has incompatible type
-            # "ExtensionDtype | dtype[Any]"; expected "dtype[Any] | type[Any]
-            # | _SupportsDType[dtype[Any]]"
-            td64_values = arr.view(dtype)  # type: ignore[arg-type]
+            td64_values = arr.view(dtype)
             return TimedeltaArray(td64_values, dtype=dtype)
 
         # error: Argument "dtype" to "view" of "_ArrayOrScalarCommon" has incompatible

pandas/core/indexes/base.py

@@ -999,17 +999,14 @@ def view(self, cls=None):
                 dtype = pandas_dtype(cls)
 
             if needs_i8_conversion(dtype):
-                if dtype.kind == "m" and dtype != "m8[ns]":
-                    # e.g. m8[s]
-                    return self._data.view(cls)
-
                 idx_cls = self._dtype_to_subclass(dtype)
-                # NB: we only get here for subclasses that override
-                #  _data_cls such that it is a type and not a tuple
-                #  of types.
-                arr_cls = idx_cls._data_cls
-                arr = arr_cls(self._data.view("i8"), dtype=dtype)
-                return idx_cls._simple_new(arr, name=self.name, refs=self._references)
+                arr = self.array.view(dtype)
+                if isinstance(arr, ExtensionArray):
+                    # here we exclude non-supported dt64/td64 dtypes
+                    return idx_cls._simple_new(
+                        arr, name=self.name, refs=self._references
+                    )
+                return arr
 
             result = self._data.view(cls)
         else:

pandas/tests/indexes/numeric/test_numeric.py

@@ -527,3 +527,19 @@ def test_map_dtype_inference_overflows():
     # TODO: we could plausibly try to infer down to int16 here
     expected = Index([1000, 2000, 3000], dtype=np.int64)
     tm.assert_index_equal(result, expected)
+
+
+def test_view_to_datetimelike():
+    # GH#55710
+    idx = Index([1, 2, 3])
+    res = idx.view("m8[s]")
+    expected = pd.TimedeltaIndex(idx.values.view("m8[s]"))
+    tm.assert_index_equal(res, expected)
+
+    res2 = idx.view("m8[D]")
+    expected2 = idx.values.view("m8[D]")
+    tm.assert_numpy_array_equal(res2, expected2)
+
+    res3 = idx.view("M8[h]")
+    expected3 = idx.values.view("M8[h]")
+    tm.assert_numpy_array_equal(res3, expected3)