BUG: Prevent OutOfBoundsDatetime error for constructing tz-aware series from list

pandas/_libs/tslib.pyi

@@ -28,5 +28,5 @@ def array_to_datetime(
 # returned ndarray may be object dtype or datetime64[ns]
 
 def array_to_datetime_with_tz(
- values: npt.NDArray[np.object_], tz: tzinfo
+ values: npt.NDArray[np.object_], tz: tzinfo, unit: str = ...
 ) -> npt.NDArray[np.int64]: ...

pandas/_libs/tslib.pyx

@@ -681,7 +681,7 @@ cdef _array_to_datetime_object(
  return oresult_nd, None
 
 
-def array_to_datetime_with_tz(ndarray values, tzinfo tz):
+def array_to_datetime_with_tz(ndarray values, tzinfo tz, unit="ns"):
  """
  Vectorized analogue to pd.Timestamp(value, tz=tz)
 
@@ -717,7 +717,7 @@ def array_to_datetime_with_tz(ndarray values, tzinfo tz):
  else:
  # datetime64, tznaive pydatetime, int, float
  ts = ts.tz_localize(tz)
- ts = ts.as_unit("ns")
+ ts = ts.as_unit(unit)
  ival = ts._value
 
  # Analogous to: result[i] = ival

pandas/core/arrays/datetimes.py

@@ -336,7 +336,7 @@ def _from_sequence_not_strict(
  # DatetimeTZDtype
  unit = dtype.unit
 
- subarr, tz, inferred_freq = _sequence_to_dt64ns(
+ subarr, tz, inferred_freq = _sequence_to_dt64(
  data,
  copy=copy,
  tz=tz,
@@ -2156,7 +2156,7 @@ def std(
 # Constructor Helpers
 
 
-def _sequence_to_dt64ns(
+def _sequence_to_dt64(
  data,
  *,
  copy: bool = False,
@@ -2182,7 +2182,8 @@ def _sequence_to_dt64ns(
  Returns
  -------
  result : numpy.ndarray
- The sequence converted to a numpy array with dtype ``datetime64[ns]``.
+ The sequence converted to a numpy array with dtype ``datetime64[unit]``.
+ Where `unit` is ns unless specified otherwise.
  tz : tzinfo or None
  Either the user-provided tzinfo or one inferred from the data.
  inferred_freq : Tick or None
@@ -2205,9 +2206,9 @@ def _sequence_to_dt64ns(
  data, copy = maybe_convert_dtype(data, copy, tz=tz)
  data_dtype = getattr(data, "dtype", None)
 
- out_dtype = DT64NS_DTYPE
- if out_unit is not None:
-  out_dtype = np.dtype(f"M8[{out_unit}]")
+ if out_unit is None:
+  out_unit = "ns"
+ out_dtype = np.dtype(f"M8[{out_unit}]")
 
  if data_dtype == object or is_string_dtype(data_dtype):
  # TODO: We do not have tests specific to string-dtypes,
@@ -2218,8 +2219,8 @@ def _sequence_to_dt64ns(
  elif tz is not None and ambiguous == "raise":
  # TODO: yearfirst/dayfirst/etc?
  obj_data = np.asarray(data, dtype=object)
- i8data = tslib.array_to_datetime_with_tz(obj_data, tz)
- return i8data.view(DT64NS_DTYPE), tz, None
+ i8data = tslib.array_to_datetime_with_tz(obj_data, tz, out_unit)
+ return i8data.view(out_dtype), tz, None
  else:
  # data comes back here as either i8 to denote UTC timestamps
  # or M8[ns] to denote wall times

pandas/tests/arrays/datetimes/test_constructors.py

@@ -8,7 +8,7 @@
 import pandas as pd
 import pandas._testing as tm
 from pandas.core.arrays import DatetimeArray
-from pandas.core.arrays.datetimes import _sequence_to_dt64ns
+from pandas.core.arrays.datetimes import _sequence_to_dt64
 
 
 class TestDatetimeArrayConstructor:
@@ -44,7 +44,7 @@ def test_freq_validation(self):
  "meth",
  [
  DatetimeArray._from_sequence,
- _sequence_to_dt64ns,
+ _sequence_to_dt64,
  pd.to_datetime,
  pd.DatetimeIndex,
  ],
@@ -105,7 +105,7 @@ def test_bool_dtype_raises(self):
  DatetimeArray._from_sequence(arr)
 
  with pytest.raises(TypeError, match=msg):
- _sequence_to_dt64ns(arr)
+ _sequence_to_dt64(arr)
 
  with pytest.raises(TypeError, match=msg):
  pd.DatetimeIndex(arr)
@@ -160,8 +160,8 @@ def test_2d(self, order):
  if order == "F":
  arr = arr.T
 
- res = _sequence_to_dt64ns(arr)
- expected = _sequence_to_dt64ns(arr.ravel())
+ res = _sequence_to_dt64(arr)
+ expected = _sequence_to_dt64(arr.ravel())
 
  tm.assert_numpy_array_equal(res[0].ravel(), expected[0])
  assert res[1] == expected[1]

pandas/tests/arrays/test_datetimelike.py

@@ -26,7 +26,7 @@
  PeriodArray,
  TimedeltaArray,
 )
-from pandas.core.arrays.datetimes import _sequence_to_dt64ns
+from pandas.core.arrays.datetimes import _sequence_to_dt64
 from pandas.core.arrays.timedeltas import sequence_to_td64ns
 
 
@@ -1313,7 +1313,7 @@ def test_from_pandas_array(dtype):
  expected = cls._from_sequence(data)
  tm.assert_extension_array_equal(result, expected)
 
- func = {"M8[ns]": _sequence_to_dt64ns, "m8[ns]": sequence_to_td64ns}[dtype]
+ func = {"M8[ns]": _sequence_to_dt64, "m8[ns]": sequence_to_td64ns}[dtype]
  result = func(arr)[0]
  expected = func(data)[0]
  tm.assert_equal(result, expected)

pandas/tests/series/test_constructors.py

@@ -1148,6 +1148,15 @@ def test_constructor_with_datetime_tz(self):
  result = DatetimeIndex(s, freq="infer")
  tm.assert_index_equal(result, dr)
 
+ def test_constructor_with_datetime_tz_ms(self):
+ # explicit frequency
+ result = Series([Timestamp("2999-01-01")], dtype="datetime64[ms, US/Pacific]")
+ expected = Series(
+ np.array(["2999-01-01"], dtype="datetime64[ms]")
+ ).dt.tz_localize("US/Pacific")
+ tm.assert_series_equal(result, expected)
+ assert result.dtype == "datetime64[ms, US/Pacific]"
+
  def test_constructor_with_datetime_tz4(self):
  # inference
  s = Series(

pandas/tests/test_downstream.py

@@ -23,7 +23,7 @@
  DatetimeArray,
  TimedeltaArray,
 )
-from pandas.core.arrays.datetimes import _sequence_to_dt64ns
+from pandas.core.arrays.datetimes import _sequence_to_dt64
 from pandas.core.arrays.timedeltas import sequence_to_td64ns
 
 
@@ -314,7 +314,7 @@ def test_from_obscure_array(dtype, array_likes):
  result = cls._from_sequence(data)
  tm.assert_extension_array_equal(result, expected)
 
- func = {"M8[ns]": _sequence_to_dt64ns, "m8[ns]": sequence_to_td64ns}[dtype]
+ func = {"M8[ns]": _sequence_to_dt64, "m8[ns]": sequence_to_td64ns}[dtype]
  result = func(arr)[0]
  expected = func(data)[0]
  tm.assert_equal(result, expected)