DEPR: tz_convert in the Timestamp constructor (pandas-dev#23621)

Author: mroeschke

doc/source/whatsnew/v0.24.0.rst

@@ -1036,6 +1036,7 @@ Deprecations
 - Constructing a :class:`TimedeltaIndex` from data with ``datetime64``-dtyped data is deprecated, will raise ``TypeError`` in a future version (:issue:`23539`)
 - The ``keep_tz=False`` option (the default) of the ``keep_tz`` keyword of
   :meth:`DatetimeIndex.to_series` is deprecated (:issue:`17832`).
+- Timezone converting a tz-aware ``datetime.datetime`` or :class:`Timestamp` with :class:`Timestamp` and the ``tz`` argument is now deprecated. Instead, use :meth:`Timestamp.tz_convert` (:issue:`23579`)
 
 .. _whatsnew_0240.deprecations.datetimelike_int_ops:
 

pandas/_libs/tslibs/timestamps.pyx

@@ -701,6 +701,9 @@ class Timestamp(_Timestamp):
             elif tz is not None:
                 raise ValueError('Can provide at most one of tz, tzinfo')
 
+            # User passed tzinfo instead of tz; avoid silently ignoring
+            tz, tzinfo = tzinfo, None
+
         if is_string_object(ts_input):
             # User passed a date string to parse.
             # Check that the user didn't also pass a date attribute kwarg.
@@ -710,24 +713,23 @@ class Timestamp(_Timestamp):
 
         elif ts_input is _no_input:
             # User passed keyword arguments.
-            if tz is None:
-                # Handle the case where the user passes `tz` and not `tzinfo`
-                tz = tzinfo
-            return Timestamp(datetime(year, month, day, hour or 0,
-                                      minute or 0, second or 0,
-                                      microsecond or 0, tzinfo),
-                             nanosecond=nanosecond, tz=tz)
+            ts_input = datetime(year, month, day, hour or 0,
+                                minute or 0, second or 0,
+                                microsecond or 0)
         elif is_integer_object(freq):
             # User passed positional arguments:
             # Timestamp(year, month, day[, hour[, minute[, second[,
             # microsecond[, nanosecond[, tzinfo]]]]]])
-            return Timestamp(datetime(ts_input, freq, tz, unit or 0,
-                                      year or 0, month or 0, day or 0,
-                                      minute), nanosecond=hour, tz=minute)
-
-        if tzinfo is not None:
-            # User passed tzinfo instead of tz; avoid silently ignoring
-            tz, tzinfo = tzinfo, None
+            ts_input = datetime(ts_input, freq, tz, unit or 0,
+                                year or 0, month or 0, day or 0)
+            nanosecond = hour
+            tz = minute
+            freq = None
+
+        if getattr(ts_input, 'tzinfo', None) is not None and tz is not None:
+            warnings.warn("Passing a datetime or Timestamp with tzinfo and the"
+                          " tz parameter will raise in the future. Use"
+                          " tz_convert instead.", FutureWarning)
 
         ts = convert_to_tsobject(ts_input, tz, unit, 0, 0, nanosecond or 0)
 

pandas/core/arrays/datetimes.py

@@ -45,7 +45,12 @@ def _to_m8(key, tz=None):
     """
     if not isinstance(key, Timestamp):
         # this also converts strings
-        key = Timestamp(key, tz=tz)
+        key = Timestamp(key)
+        if key.tzinfo is not None and tz is not None:
+            # Don't tz_localize(None) if key is already tz-aware
+            key = key.tz_convert(tz)
+        else:
+            key = key.tz_localize(tz)
 
     return np.int64(conversion.pydt_to_i8(key)).view(_NS_DTYPE)
 

pandas/core/generic.py

@@ -9336,8 +9336,14 @@ def describe_categorical_1d(data):
                 if is_datetime64_any_dtype(data):
                     tz = data.dt.tz
                     asint = data.dropna().values.view('i8')
+                    top = Timestamp(top)
+                    if top.tzinfo is not None and tz is not None:
+                        # Don't tz_localize(None) if key is already tz-aware
+                        top = top.tz_convert(tz)
+                    else:
+                        top = top.tz_localize(tz)
                     names += ['top', 'freq', 'first', 'last']
-                    result += [Timestamp(top, tz=tz), freq,
+                    result += [top, freq,
                                Timestamp(asint.min(), tz=tz),
                                Timestamp(asint.max(), tz=tz)]
                 else:

pandas/core/indexes/datetimes.py

@@ -937,7 +937,10 @@ def get_value(self, series, key):
 
             # needed to localize naive datetimes
             if self.tz is not None:
-                key = Timestamp(key, tz=self.tz)
+                if key.tzinfo is not None:
+                    key = Timestamp(key).tz_convert(self.tz)
+                else:
+                    key = Timestamp(key).tz_localize(self.tz)
 
             return self.get_value_maybe_box(series, key)
 
@@ -963,7 +966,11 @@ def get_value(self, series, key):
     def get_value_maybe_box(self, series, key):
         # needed to localize naive datetimes
         if self.tz is not None:
-            key = Timestamp(key, tz=self.tz)
+            key = Timestamp(key)
+            if key.tzinfo is not None:
+                key = key.tz_convert(self.tz)
+            else:
+                key = key.tz_localize(self.tz)
         elif not isinstance(key, Timestamp):
             key = Timestamp(key)
         values = self._engine.get_value(com.values_from_object(series),
@@ -986,7 +993,10 @@ def get_loc(self, key, method=None, tolerance=None):
 
         if isinstance(key, datetime):
             # needed to localize naive datetimes
-            key = Timestamp(key, tz=self.tz)
+            if key.tzinfo is None:
+                key = Timestamp(key, tz=self.tz)
+            else:
+                key = Timestamp(key).tz_convert(self.tz)
             return Index.get_loc(self, key, method, tolerance)
 
         elif isinstance(key, timedelta):
@@ -1010,7 +1020,11 @@ def get_loc(self, key, method=None, tolerance=None):
                 pass
 
             try:
-                stamp = Timestamp(key, tz=self.tz)
+                stamp = Timestamp(key)
+                if stamp.tzinfo is not None and self.tz is not None:
+                    stamp = stamp.tz_convert(self.tz)
+                else:
+                    stamp = stamp.tz_localize(self.tz)
                 return Index.get_loc(self, stamp, method, tolerance)
             except KeyError:
                 raise KeyError(key)

pandas/io/formats/format.py

@@ -1246,7 +1246,10 @@ def _format_datetime64(x, tz=None, nat_rep='NaT'):
         return nat_rep
 
     if tz is not None or not isinstance(x, Timestamp):
-        x = Timestamp(x, tz=tz)
+        if getattr(x, 'tzinfo', None) is not None:
+            x = Timestamp(x).tz_convert(tz)
+        else:
+            x = Timestamp(x).tz_localize(tz)
 
     return str(x)
 

pandas/tests/indexes/datetimes/test_construction.py

@@ -521,6 +521,12 @@ def test_construction_from_replaced_timestamps_with_dst(self):
                                     tz='Australia/Melbourne')
         tm.assert_index_equal(result, expected)
 
+    def test_construction_with_tz_and_tz_aware_dti(self):
+        # GH 23579
+        dti = date_range('2016-01-01', periods=3, tz='US/Central')
+        with pytest.raises(TypeError):
+            DatetimeIndex(dti, tz='Asia/Tokyo')
+
 
 class TestTimeSeries(object):
 

pandas/tests/scalar/timestamp/test_timestamp.py

@@ -244,7 +244,10 @@ def test_constructor(self):
                     assert conversion.pydt_to_i8(result) == expected_tz
 
                     # should convert to UTC
-                    result = Timestamp(result, tz='UTC')
+                    if tz is not None:
+                        result = Timestamp(result).tz_convert('UTC')
+                    else:
+                        result = Timestamp(result, tz='UTC')
                     expected_utc = expected - offset * 3600 * 1000000000
                     assert result.value == expected_utc
                     assert conversion.pydt_to_i8(result) == expected_utc
@@ -295,7 +298,7 @@ def test_constructor_with_stringoffset(self):
                 assert conversion.pydt_to_i8(result) == expected_tz
 
                 # should convert to UTC
-                result = Timestamp(result, tz='UTC')
+                result = Timestamp(result).tz_convert('UTC')
                 expected_utc = expected
                 assert result.value == expected_utc
                 assert conversion.pydt_to_i8(result) == expected_utc
@@ -558,7 +561,7 @@ def test_construct_timestamp_near_dst(self, offset):
         # GH 20854
         expected = Timestamp('2016-10-30 03:00:00{}'.format(offset),
                              tz='Europe/Helsinki')
-        result = Timestamp(expected, tz='Europe/Helsinki')
+        result = Timestamp(expected).tz_convert('Europe/Helsinki')
         assert result == expected
 
     @pytest.mark.parametrize('arg', [
@@ -580,6 +583,13 @@ def test_constructor_invalid_frequency(self):
         with pytest.raises(ValueError, match="Invalid frequency:"):
             Timestamp('2012-01-01', freq=[])
 
+    @pytest.mark.parametrize('box', [datetime, Timestamp])
+    def test_depreciate_tz_and_tzinfo_in_datetime_input(self, box):
+        # GH 23579
+        kwargs = {'year': 2018, 'month': 1, 'day': 1, 'tzinfo': utc}
+        with tm.assert_produces_warning(FutureWarning):
+            Timestamp(box(**kwargs), tz='US/Pacific')
+
 
 class TestTimestamp(object):