BUG: cannot subtract Timestamp with different timezones (pandas-dev#31793)

Author: AnjoMan

doc/source/whatsnew/v1.3.0.rst

@@ -936,6 +936,7 @@ Timezones
 ^^^^^^^^^
 - Bug in different ``tzinfo`` objects representing UTC not being treated as equivalent (:issue:`39216`)
 - Bug in ``dateutil.tz.gettz("UTC")`` not being recognized as equivalent to other UTC-representing tzinfos (:issue:`39276`)
+- Bug in ``Timestamp`` and ``DatetimeIndex`` incorrectly raising a ``TypeError`` when subtracting two times that don't have the same timezone offset (:issue:`31793`)
 
 Numeric
 ^^^^^^^

pandas/_libs/tslibs/timestamps.pyx

@@ -343,10 +343,10 @@ cdef class _Timestamp(ABCTimestamp):
             else:
                 self = type(other)(self)
 
-            # validate tz's
-            if not tz_compare(self.tzinfo, other.tzinfo):
-                raise TypeError("Timestamp subtraction must have the "
-                                "same timezones or no timezones")
+            if (self.tzinfo is None) ^ (other.tzinfo is None):
+                raise TypeError(
+                    "Cannot compare tz-naive and tz-aware datetime-like objects."
+                )
 
             # scalar Timestamp/datetime - Timestamp/datetime -> yields a
             # Timedelta

pandas/core/arrays/datetimes.py

@@ -708,12 +708,7 @@ def _sub_datetime_arraylike(self, other):
             assert is_datetime64_dtype(other)
             other = type(self)(other)
 
-        if not self._has_same_tz(other):
-            # require tz compat
-            raise TypeError(
-                f"{type(self).__name__} subtraction must have the same "
-                "timezones or no timezones"
-            )
+        self._assert_tzawareness_compat(other)
 
         self_i8 = self.asi8
         other_i8 = other.asi8
@@ -759,11 +754,7 @@ def _sub_datetimelike_scalar(self, other):
         if other is NaT:  # type: ignore[comparison-overlap]
             return self - NaT
 
-        if not self._has_same_tz(other):
-            # require tz compat
-            raise TypeError(
-                "Timestamp subtraction must have the same timezones or no timezones"
-            )
+        self._assert_tzawareness_compat(other)
 
         i8 = self.asi8
         result = checked_add_with_arr(i8, -other.value, arr_mask=self._isnan)

pandas/tests/arithmetic/test_datetime64.py

@@ -864,6 +864,46 @@ def test_dt64arr_isub_timedeltalike_scalar(
         rng -= two_hours
         tm.assert_equal(rng, expected)
 
+    def test_dt64_array_sub_dt_with_different_timezone(self, box_with_array):
+        t1 = date_range("20130101", periods=3).tz_localize("US/Eastern")
+        t1 = tm.box_expected(t1, box_with_array)
+        t2 = Timestamp("20130101").tz_localize("CET")
+
+        result = t1 - t2
+        expected = TimedeltaIndex(
+            ["0 days 06:00:00", "1 days 06:00:00", "2 days 06:00:00"]
+        )
+        expected = tm.box_expected(expected, box_with_array)
+        tm.assert_equal(result, expected)
+
+        result = t2 - t1
+        expected = TimedeltaIndex(
+            ["-1 days +18:00:00", "-2 days +18:00:00", "-3 days +18:00:00"]
+        )
+        expected = tm.box_expected(expected, box_with_array)
+        tm.assert_equal(result, expected)
+
+    def test_dt64_array_sub_dt64_array_with_different_timezone(self, box_with_array):
+        t1 = date_range("20130101", periods=3).tz_localize("US/Eastern")
+        t1 = tm.box_expected(t1, box_with_array)
+
+        t2 = date_range("20130101", periods=3).tz_localize("CET")
+        t2 = tm.box_expected(t2, box_with_array)
+
+        result = t1 - t2
+        expected = TimedeltaIndex(
+            ["0 days 06:00:00", "0 days 06:00:00", "0 days 06:00:00"]
+        )
+        expected = tm.box_expected(expected, box_with_array)
+        tm.assert_equal(result, expected)
+
+        result = t2 - t1
+        expected = TimedeltaIndex(
+            ["-1 days +18:00:00", "-1 days +18:00:00", "-1 days +18:00:00"]
+        )
+        expected = tm.box_expected(expected, box_with_array)
+        tm.assert_equal(result, expected)
+
     # TODO: redundant with test_dt64arr_add_timedeltalike_scalar
     def test_dt64arr_add_td64_scalar(self, box_with_array):
         # scalar timedeltas/np.timedelta64 objects
@@ -1045,7 +1085,7 @@ def test_dt64arr_aware_sub_dt64ndarray_raises(
         dt64vals = dti.values
 
         dtarr = tm.box_expected(dti, box_with_array)
-        msg = "subtraction must have the same timezones or"
+        msg = "Cannot compare tz-naive and tz-aware datetime"
         with pytest.raises(TypeError, match=msg):
             dtarr - dt64vals
         with pytest.raises(TypeError, match=msg):
@@ -2235,24 +2275,20 @@ def test_sub_dti_dti(self):
 
         dti = date_range("20130101", periods=3)
         dti_tz = date_range("20130101", periods=3).tz_localize("US/Eastern")
-        dti_tz2 = date_range("20130101", periods=3).tz_localize("UTC")
         expected = TimedeltaIndex([0, 0, 0])
 
         result = dti - dti
         tm.assert_index_equal(result, expected)
 
         result = dti_tz - dti_tz
         tm.assert_index_equal(result, expected)
-        msg = "DatetimeArray subtraction must have the same timezones or"
+        msg = "Cannot compare tz-naive and tz-aware datetime-like objects"
         with pytest.raises(TypeError, match=msg):
             dti_tz - dti
 
         with pytest.raises(TypeError, match=msg):
             dti - dti_tz
 
-        with pytest.raises(TypeError, match=msg):
-            dti_tz - dti_tz2
-
         # isub
         dti -= dti
         tm.assert_index_equal(dti, expected)

pandas/tests/arithmetic/test_timedelta64.py

@@ -384,36 +384,31 @@ def _check(result, expected):
         expected = Timedelta("0 days")
         _check(result, expected)
 
+        # gotta chedk this
         # tz mismatches
-        msg = "Timestamp subtraction must have the same timezones or no timezones"
+        msg = "Cannot compare tz-naive and tz-aware datetime-like objects."
         with pytest.raises(TypeError, match=msg):
             dt_tz - ts
         msg = "can't subtract offset-naive and offset-aware datetimes"
         with pytest.raises(TypeError, match=msg):
             dt_tz - dt
-        msg = "Timestamp subtraction must have the same timezones or no timezones"
-        with pytest.raises(TypeError, match=msg):
-            dt_tz - ts_tz2
         msg = "can't subtract offset-naive and offset-aware datetimes"
         with pytest.raises(TypeError, match=msg):
             dt - dt_tz
-        msg = "Timestamp subtraction must have the same timezones or no timezones"
+        msg = "Cannot compare tz-naive and tz-aware datetime-like objects."
         with pytest.raises(TypeError, match=msg):
             ts - dt_tz
         with pytest.raises(TypeError, match=msg):
             ts_tz2 - ts
         with pytest.raises(TypeError, match=msg):
             ts_tz2 - dt
-        with pytest.raises(TypeError, match=msg):
-            ts_tz - ts_tz2
 
+        msg = "Cannot compare tz-naive and tz-aware"
         # with dti
         with pytest.raises(TypeError, match=msg):
             dti - ts_tz
         with pytest.raises(TypeError, match=msg):
             dti_tz - ts
-        with pytest.raises(TypeError, match=msg):
-            dti_tz - ts_tz2
 
         result = dti_tz - dt_tz
         expected = TimedeltaIndex(["0 days", "1 days", "2 days"])

pandas/tests/scalar/timestamp/test_arithmetic.py

@@ -1,6 +1,7 @@
 from datetime import (
     datetime,
     timedelta,
+    timezone,
 )
 
 import numpy as np
@@ -99,7 +100,7 @@ def test_rsub_dtscalars(self, tz_naive_fixture):
         if tz_naive_fixture is None:
             assert other.to_datetime64() - ts == td
         else:
-            msg = "subtraction must have"
+            msg = "Cannot compare tz-naive and tz-aware datetime-like objects"
             with pytest.raises(TypeError, match=msg):
                 other.to_datetime64() - ts
 
@@ -109,6 +110,37 @@ def test_timestamp_sub_datetime(self):
         assert (ts - dt).days == 1
         assert (dt - ts).days == -1
 
+    def test_subtract_tzaware_datetime(self):
+        t1 = Timestamp("2020-10-22T22:00:00+00:00")
+        t2 = datetime(2020, 10, 22, 22, tzinfo=timezone.utc)
+
+        result = t1 - t2
+
+        assert isinstance(result, Timedelta)
+        assert result == Timedelta("0 days")
+
+    def test_subtract_timestamp_from_different_timezone(self):
+        t1 = Timestamp("20130101").tz_localize("US/Eastern")
+        t2 = Timestamp("20130101").tz_localize("CET")
+
+        result = t1 - t2
+
+        assert isinstance(result, Timedelta)
+        assert result == Timedelta("0 days 06:00:00")
+
+    def test_subtracting_involving_datetime_with_different_tz(self):
+        t1 = datetime(2013, 1, 1, tzinfo=timezone(timedelta(hours=-5)))
+        t2 = Timestamp("20130101").tz_localize("CET")
+
+        result = t1 - t2
+
+        assert isinstance(result, Timedelta)
+        assert result == Timedelta("0 days 06:00:00")
+
+        result = t2 - t1
+        assert isinstance(result, Timedelta)
+        assert result == Timedelta("-1 days +18:00:00")
+
     def test_addition_subtraction_types(self):
         # Assert on the types resulting from Timestamp +/- various date/time
         # objects

pandas/tests/series/methods/test_shift.py

@@ -110,11 +110,10 @@ def test_shift(self, datetime_series):
         exp = Series(TimedeltaIndex(["NaT"] + ["1 days"] * 4), name="foo")
         tm.assert_series_equal(result, exp)
 
-        # incompat tz
+        # mismatched timezone
         s2 = Series(date_range("2000-01-01 09:00:00", periods=5, tz="CET"), name="foo")
-        msg = "DatetimeArray subtraction must have the same timezones or no timezones"
-        with pytest.raises(TypeError, match=msg):
-            s - s2
+        exp = Series(TimedeltaIndex(["6 hours"] * 5), name="foo")
+        result = s - s2
 
     def test_shift2(self):
         ts = Series(