BUG: Fix localize_pydatetime using meta datetimes as Timestamp (#25734)

doc/source/whatsnew/v0.25.0.rst

@@ -167,7 +167,7 @@ Numeric
 Conversion
 ^^^^^^^^^^
 
--
+- Bug in :meth:`localize_pydatetime` where some ``datetime`` were incorrectly treated as pandas :class:`Timestamp` (:issue:`25734`)
 -
 -
 

pandas/_libs/tslibs/conversion.pyx

@@ -27,7 +27,8 @@ from pandas._libs.tslibs.np_datetime cimport (
 from pandas._libs.tslibs.np_datetime import OutOfBoundsDatetime
 
 from pandas._libs.tslibs.util cimport (
-    is_string_object, is_datetime64_object, is_integer_object, is_float_object)
+    is_string_object, is_datetime64_object, is_integer_object, is_float_object,
+    is_timestamp)
 
 from pandas._libs.tslibs.timedeltas cimport (cast_from_unit,
                                              delta_to_nanoseconds)
@@ -607,8 +608,7 @@ cpdef inline datetime localize_pydatetime(datetime dt, object tz):
     """
     if tz is None:
         return dt
-    elif not PyDateTime_CheckExact(dt):
-        # i.e. is a Timestamp
+    elif is_timestamp(dt):
         return dt.tz_localize(tz)
     elif is_utc(tz):
         return _localize_pydatetime(dt, tz)

pandas/_libs/tslibs/timestamps.pyx

@@ -203,6 +203,8 @@ def round_nsint64(values, mode, freq):
 # shadows the python class, where we do any heavy lifting.
 cdef class _Timestamp(datetime):
 
+    _typ = 'timestamp'
+
     cdef readonly:
         int64_t value, nanosecond
         object freq       # frequency reference

pandas/_libs/tslibs/util.pxd

@@ -227,3 +227,18 @@ cdef inline bint is_nan(object val):
     is_nan : bool
     """
     return (is_float_object(val) or is_complex_object(val)) and val != val
+
+
+cdef inline bint is_timestamp(object val):
+    """
+    Cython equivalent of `isinstance(val, pd.Timestamp)`
+
+    Parameters
+    ----------
+    val : object
+
+    Returns
+    -------
+    is_timestamp : bool
+    """
+    return getattr(val, '_typ', None) == "timestamp"

pandas/tests/tslibs/test_conversion.py

@@ -3,13 +3,16 @@
 import numpy as np
 import pytest
 from pytz import UTC
+from datetime import datetime
 
 from pandas._libs.tslib import iNaT
 from pandas._libs.tslibs import conversion, timezones
 
 from pandas import date_range
 import pandas.util.testing as tm
 
+from pandas import Timestamp
+
 
 def _compare_utc_to_local(tz_didx):
     def f(x):
@@ -66,3 +69,26 @@ def test_length_zero_copy(dtype, copy):
     arr = np.array([], dtype=dtype)
     result = conversion.ensure_datetime64ns(arr, copy=copy)
     assert result.base is (None if copy else arr)
+
+
+class MetaDatetime(type):
+    pass
+
+
+class FakeDatetime(MetaDatetime("NewBase", (datetime,), {})):
+    pass
+
+
+@pytest.mark.parametrize("dt, expected", [
+    pytest.param(Timestamp("2000-01-01"),
+                 Timestamp("2000-01-01", tz=UTC), id="timestamp"),
+    pytest.param(datetime(2000, 1, 1),
+                 datetime(2000, 1, 1, tzinfo=UTC),
+                 id="datetime"),
+    pytest.param(FakeDatetime(2000, 1, 1),
+                 FakeDatetime(2000, 1, 1, tzinfo=UTC),
+                 id="fakedatetime")])
+def test_localize_pydatetime_dt_types(dt, expected):
+    # GH 25734
+    result = conversion.localize_pydatetime(dt, UTC)
+    assert result == expected