BUG: DataFrame(dt64data, dtype=td64) corner cases (#38792)

Author: jbrockmendel

doc/source/whatsnew/v1.3.0.rst

@@ -190,7 +190,8 @@ Datetimelike
 ^^^^^^^^^^^^
 - Bug in :class:`DataFrame` and :class:`Series` constructors sometimes dropping nanoseconds from :class:`Timestamp` (resp. :class:`Timedelta`) ``data``, with ``dtype=datetime64[ns]`` (resp. ``timedelta64[ns]``) (:issue:`38032`)
 - Bug in :meth:`DataFrame.first` and :meth:`Series.first` returning two months for offset one month when first day is last calendar day (:issue:`29623`)
-- Bug in constructing a :class:`DataFrame` or :class:`Series` with mismatched ``datetime64`` data and ``timedelta64`` dtype, or vice-versa, failing to raise ``TypeError`` (:issue:`38575`, :issue:`38764`)
+- Bug in constructing a :class:`DataFrame` or :class:`Series` with mismatched ``datetime64`` data and ``timedelta64`` dtype, or vice-versa, failing to raise ``TypeError`` (:issue:`38575`, :issue:`38764`, :issue:`38792`)
+- Bug in constructing a :class:`Series` or :class:`DataFrame` with a ``datetime`` object out of bounds for ``datetime64[ns]`` dtype (:issue:`38792`)
 - Bug in :meth:`DatetimeIndex.intersection`, :meth:`DatetimeIndex.symmetric_difference`, :meth:`PeriodIndex.intersection`, :meth:`PeriodIndex.symmetric_difference` always returning object-dtype when operating with :class:`CategoricalIndex` (:issue:`38741`)
 - Bug in :meth:`Series.where` incorrectly casting ``datetime64`` values to ``int64`` (:issue:`37682`)
 -

pandas/core/construction.py

@@ -615,9 +615,12 @@ def _try_cast(arr, dtype: Optional[DtypeObj], copy: bool, raise_cast_failure: bo
     except OutOfBoundsDatetime:
         # in case of out of bound datetime64 -> always raise
         raise
-    except (ValueError, TypeError):
+    except (ValueError, TypeError) as err:
         if dtype is not None and raise_cast_failure:
             raise
+        elif "Cannot cast" in str(err):
+            # via _disallow_mismatched_datetimelike
+            raise
         else:
             subarr = np.array(arr, dtype=object, copy=copy)
     return subarr

pandas/core/dtypes/cast.py

@@ -170,7 +170,7 @@ def maybe_unbox_datetimelike(value: Scalar, dtype: DtypeObj) -> Scalar:
     return value
 
 
-def _disallow_mismatched_datetimelike(value: DtypeObj, dtype: DtypeObj):
+def _disallow_mismatched_datetimelike(value, dtype: DtypeObj):
     """
     numpy allows np.array(dt64values, dtype="timedelta64[ns]") and
     vice-versa, but we do not want to allow this, so we need to
@@ -725,7 +725,11 @@ def infer_dtype_from_scalar(val, pandas_dtype: bool = False) -> Tuple[DtypeObj,
         dtype = np.dtype(object)
 
     elif isinstance(val, (np.datetime64, datetime)):
-        val = Timestamp(val)
+        try:
+            val = Timestamp(val)
+        except OutOfBoundsDatetime:
+            return np.dtype(object), val
+
         if val is NaT or val.tz is None:
             dtype = np.dtype("M8[ns]")
         else:
@@ -1472,6 +1476,8 @@ def maybe_cast_to_datetime(value, dtype: Optional[DtypeObj]):
 
                 # we have an array of datetime or timedeltas & nulls
                 elif np.prod(value.shape) or not is_dtype_equal(value.dtype, dtype):
+                    _disallow_mismatched_datetimelike(value, dtype)
+
                     try:
                         if is_datetime64:
                             value = to_datetime(value, errors="raise")

pandas/tests/frame/test_constructors.py

@@ -2987,12 +2987,11 @@ def test_from_timedelta64_scalar_object(self, constructor, request):
     def test_from_scalar_datetimelike_mismatched(self, constructor, cls, request):
         node = request.node
         params = node.callspec.params
-        if params["frame_or_series"] is DataFrame and params["constructor"] is not None:
+        if params["frame_or_series"] is DataFrame and params["constructor"] is dict:
             mark = pytest.mark.xfail(
                 reason="DataFrame incorrectly allows mismatched datetimelike"
             )
             node.add_marker(mark)
-
         scalar = cls("NaT", "ns")
         dtype = {np.datetime64: "m8[ns]", np.timedelta64: "M8[ns]"}[cls]
 
@@ -3002,3 +3001,9 @@ def test_from_scalar_datetimelike_mismatched(self, constructor, cls, request):
         scalar = cls(4, "ns")
         with pytest.raises(TypeError, match="Cannot cast"):
             constructor(scalar, dtype=dtype)
+
+    def test_from_out_of_bounds_datetime(self, constructor):
+        scalar = datetime(9999, 1, 1)
+        result = constructor(scalar)
+
+        assert type(get1(result)) is datetime