BUG: pd.to_datetime([np.str_ objects]) GH#32264

doc/source/whatsnew/v1.5.0.rst

@@ -118,7 +118,7 @@ Categorical
 
 Datetimelike
 ^^^^^^^^^^^^
--
+- Bug in :func:`to_datetime` with sequences of ``np.str_`` objects incorrectly raising (:issue:`32264`)
 -
 
 Timedelta

pandas/_libs/tslib.pyx

@@ -506,6 +506,9 @@ cpdef array_to_datetime(
                 elif isinstance(val, str):
                     # string
                     seen_string = True
+                    if type(val) is not str:
+                        # GH#32264 np.str_ object
+                        val = str(val)
 
                     if len(val) == 0 or val in nat_strings:
                         iresult[i] = NPY_NAT
@@ -735,6 +738,10 @@ cdef _array_to_datetime_object(
             # GH 25978. No need to parse NaT-like or datetime-like vals
             oresult[i] = val
         elif isinstance(val, str):
+            if type(val) is not str:
+                # GH#32264 np.str_ objects
+                val = str(val)
+
             if len(val) == 0 or val in nat_strings:
                 oresult[i] = 'NaT'
                 continue

pandas/_libs/tslibs/parsing.pyx

@@ -239,6 +239,9 @@ cdef inline bint does_string_look_like_time(str parse_string):
 
 
 def parse_datetime_string(
+    # NB: This will break with np.str_ (GH#32264) even though
+    #  isinstance(npstrobj, str) evaluates to True, so caller must ensure
+    #  the argument is *exactly* 'str'
     str date_string,
     bint dayfirst=False,
     bint yearfirst=False,
@@ -254,7 +257,7 @@ def parse_datetime_string(
     """
 
     cdef:
-        object dt
+        datetime dt
 
     if not _does_string_look_like_datetime(date_string):
         raise ValueError('Given date string not likely a datetime.')

pandas/tests/tools/test_to_datetime.py

@@ -455,6 +455,25 @@ def test_to_datetime_parse_timezone_keeps_name(self):
 
 
 class TestToDatetime:
+    def test_to_datetime_np_str(self):
+        # GH#32264
+        value = np.str_("2019-02-04 10:18:46.297000+0000")
+
+        ser = Series([value])
+
+        exp = Timestamp("2019-02-04 10:18:46.297000", tz="UTC")
+
+        assert to_datetime(value) == exp
+        assert to_datetime(ser.iloc[0]) == exp
+
+        res = to_datetime([value])
+        expected = Index([exp])
+        tm.assert_index_equal(res, expected)
+
+        res = to_datetime(ser)
+        expected = Series(expected)
+        tm.assert_series_equal(res, expected)
+
     @pytest.mark.parametrize(
         "s, _format, dt",
         [