API: default to stdlib timezone objects for fixed-offsets

doc/source/whatsnew/v2.0.0.rst

@@ -325,6 +325,7 @@ Other API changes
 ^^^^^^^^^^^^^^^^^
 - The ``freq``, ``tz``, ``nanosecond``, and ``unit`` keywords in the :class:`Timestamp` constructor are now keyword-only (:issue:`45307`)
 - Passing ``nanoseconds`` greater than 999 or less than 0 in :class:`Timestamp` now raises a ``ValueError`` (:issue:`48538`, :issue:`48255`)
+- When inferring a fixed-offset ``tzinfo`` from a string or integer ``tz``, a standard-library ``datetime.timezone`` object is returned instead of a ``pytz`` object (:issue:`34916`)
 - :func:`read_csv`: specifying an incorrect number of columns with ``index_col`` of now raises ``ParserError`` instead of ``IndexError`` when using the c parser.
 - Default value of ``dtype`` in :func:`get_dummies` is changed to ``bool`` from ``uint8`` (:issue:`45848`)
 - :meth:`DataFrame.astype`, :meth:`Series.astype`, and :meth:`DatetimeIndex.astype` casting datetime64 data to any of "datetime64[s]", "datetime64[ms]", "datetime64[us]" will return an object with the given resolution instead of coercing back to "datetime64[ns]" (:issue:`48928`)

pandas/_libs/tslib.pyx

@@ -1,9 +1,13 @@
 cimport cython
+
+from datetime import timezone
+
 from cpython.datetime cimport (
     PyDate_Check,
     PyDateTime_Check,
     datetime,
     import_datetime,
+    timedelta,
     tzinfo,
 )
 from cpython.object cimport PyObject
@@ -23,8 +27,6 @@ import numpy as np
 
 cnp.import_array()
 
-import pytz
-
 from pandas._libs.tslibs.np_datetime cimport (
     NPY_DATETIMEUNIT,
     NPY_FR_ns,
@@ -94,7 +96,7 @@ def _test_parse_iso8601(ts: str):
     obj.value = npy_datetimestruct_to_datetime(NPY_FR_ns, &obj.dts)
     check_dts_bounds(&obj.dts)
     if out_local == 1:
-        obj.tzinfo = pytz.FixedOffset(out_tzoffset)
+        obj.tzinfo = timezone(timedelta(minutes=out_tzoffset))
         obj.value = tz_localize_to_utc_single(obj.value, obj.tzinfo)
         return Timestamp(obj.value, tz=obj.tzinfo)
     else:
@@ -454,7 +456,7 @@ cpdef array_to_datetime(
         2) datetime.datetime objects, if OutOfBoundsDatetime or TypeError
            is encountered
 
-    Also returns a pytz.FixedOffset if an array of strings with the same
+    Also returns a fixed-offset tzinfo object if an array of strings with the same
     timezone offset is passed and utc=True is not passed. Otherwise, None
     is returned
 
@@ -655,7 +657,7 @@ cpdef array_to_datetime(
                             # since we store the total_seconds of
                             # dateutil.tz.tzoffset objects
                             out_tzoffset_vals.add(out_tzoffset * 60.)
-                            tz = pytz.FixedOffset(out_tzoffset)
+                            tz = timezone(timedelta(minutes=out_tzoffset))
                             value = tz_localize_to_utc_single(value, tz)
                             out_local = 0
                             out_tzoffset = 0
@@ -729,7 +731,7 @@ cpdef array_to_datetime(
             return _array_to_datetime_object(values, errors, dayfirst, yearfirst)
         else:
             tz_offset = out_tzoffset_vals.pop()
-            tz_out = pytz.FixedOffset(tz_offset / 60.)
+            tz_out = timezone(timedelta(seconds=tz_offset))
     return result, tz_out
 
 

pandas/_libs/tslibs/conversion.pyx

@@ -8,16 +8,17 @@ from numpy cimport (
 
 cnp.import_array()
 
-import pytz
-
 # stdlib datetime imports
 
+from datetime import timezone
+
 from cpython.datetime cimport (
     PyDate_Check,
     PyDateTime_Check,
     datetime,
     import_datetime,
     time,
+    timedelta,
     tzinfo,
 )
 
@@ -407,7 +408,7 @@ cdef _TSObject _create_tsobject_tz_using_offset(npy_datetimestruct dts,
 
     value = npy_datetimestruct_to_datetime(NPY_FR_ns, &dts)
     obj.dts = dts
-    obj.tzinfo = pytz.FixedOffset(tzoffset)
+    obj.tzinfo = timezone(timedelta(minutes=tzoffset))
     obj.value = tz_localize_to_utc_single(value, obj.tzinfo)
     if tz is None:
         check_overflows(obj, NPY_FR_ns)

pandas/_libs/tslibs/strptime.pyx

@@ -1,7 +1,10 @@
 """Strptime-related classes and functions.
 """
+from datetime import timezone
+
 from cpython.datetime cimport (
     date,
+    timedelta,
     tzinfo,
 )
 
@@ -488,15 +491,15 @@ cdef (int, int) _calc_julian_from_V(int iso_year, int iso_week, int iso_weekday)
 
 cdef tzinfo parse_timezone_directive(str z):
     """
-    Parse the '%z' directive and return a pytz.FixedOffset
+    Parse the '%z' directive and return a datetime.timezone object.
 
     Parameters
     ----------
     z : string of the UTC offset
 
     Returns
     -------
-    pytz.FixedOffset
+    datetime.timezone
 
     Notes
     -----
@@ -510,7 +513,7 @@ cdef tzinfo parse_timezone_directive(str z):
         object gmtoff_remainder, gmtoff_remainder_padding
 
     if z == 'Z':
-        return pytz.FixedOffset(0)
+        return timezone(timedelta(0))
     if z[3] == ':':
         z = z[:3] + z[4:]
         if len(z) > 5:
@@ -530,4 +533,4 @@ cdef tzinfo parse_timezone_directive(str z):
     total_minutes = ((hours * 60) + minutes + (seconds // 60) +
                      (microseconds // 60_000_000))
     total_minutes = -total_minutes if z.startswith("-") else total_minutes
-    return pytz.FixedOffset(total_minutes)
+    return timezone(timedelta(minutes=total_minutes))

pandas/_libs/tslibs/timestamps.pyx

@@ -111,7 +111,7 @@ from pandas._libs.tslibs.timezones cimport (
     is_utc,
     maybe_get_tz,
     treat_tz_as_pytz,
-    utc_pytz as UTC,
+    utc_stdlib as UTC,
 )
 from pandas._libs.tslibs.tzconversion cimport (
     tz_convert_from_utc_single,

pandas/_libs/tslibs/timezones.pxd

@@ -5,7 +5,7 @@ from cpython.datetime cimport (
 )
 
 
-cdef tzinfo utc_pytz
+cdef tzinfo utc_stdlib
 
 cpdef bint is_utc(tzinfo tz)
 cdef bint is_tzlocal(tzinfo tz)

pandas/_libs/tslibs/timezones.pyx

@@ -27,14 +27,10 @@ from dateutil.tz import (
     tzlocal as _dateutil_tzlocal,
     tzutc as _dateutil_tzutc,
 )
+import numpy as np
 import pytz
 from pytz.tzinfo import BaseTzInfo as _pytz_BaseTzInfo
 
-UTC = pytz.utc
-
-
-import numpy as np
-
 cimport numpy as cnp
 from numpy cimport int64_t
 
@@ -49,7 +45,7 @@ from pandas._libs.tslibs.util cimport (
 
 cdef int64_t NPY_NAT = get_nat()
 cdef tzinfo utc_stdlib = timezone.utc
-cdef tzinfo utc_pytz = UTC
+cdef tzinfo utc_pytz = pytz.utc
 cdef tzinfo utc_dateutil_str = dateutil_gettz("UTC")  # NB: *not* the same as tzutc()
 
 cdef tzinfo utc_zoneinfo = None
@@ -171,7 +167,7 @@ cpdef inline tzinfo maybe_get_tz(object tz):
         else:
             tz = pytz.timezone(tz)
     elif is_integer_object(tz):
-        tz = pytz.FixedOffset(tz / 60)
+        tz = timezone(timedelta(seconds=tz))
     elif isinstance(tz, tzinfo):
         pass
     elif tz is None:

pandas/core/tools/datetimes.py

@@ -991,7 +991,7 @@ def to_datetime(
 
     >>> pd.to_datetime(['2018-10-26 12:00 -0500', '2018-10-26 13:00 -0500'])
     DatetimeIndex(['2018-10-26 12:00:00-05:00', '2018-10-26 13:00:00-05:00'],
-                  dtype='datetime64[ns, pytz.FixedOffset(-300)]', freq=None)
+                  dtype='datetime64[ns, UTC-05:00]', freq=None)
 
     - However, timezone-aware inputs *with mixed time offsets* (for example
       issued from a timezone with daylight savings, such as Europe/Paris)
@@ -1010,7 +1010,7 @@ def to_datetime(
     >>> from datetime import datetime
     >>> pd.to_datetime(["2020-01-01 01:00 -01:00", datetime(2020, 1, 1, 3, 0)])
     DatetimeIndex(['2020-01-01 01:00:00-01:00', '2020-01-01 02:00:00-01:00'],
-                  dtype='datetime64[ns, pytz.FixedOffset(-60)]', freq=None)
+                  dtype='datetime64[ns, UTC-01:00]', freq=None)
 
     |
 

pandas/tests/indexes/datetimes/test_constructors.py

@@ -627,15 +627,15 @@ def test_constructor_coverage(self):
 
     @pytest.mark.parametrize("freq", ["AS", "W-SUN"])
     def test_constructor_datetime64_tzformat(self, freq):
-        # see GH#6572: ISO 8601 format results in pytz.FixedOffset
+        # see GH#6572: ISO 8601 format results in stdlib timezone object
         idx = date_range(
             "2013-01-01T00:00:00-05:00", "2016-01-01T23:59:59-05:00", freq=freq
         )
         expected = date_range(
             "2013-01-01T00:00:00",
             "2016-01-01T23:59:59",
             freq=freq,
-            tz=pytz.FixedOffset(-300),
+            tz=timezone(timedelta(minutes=-300)),
         )
         tm.assert_index_equal(idx, expected)
         # Unable to use `US/Eastern` because of DST
@@ -651,7 +651,7 @@ def test_constructor_datetime64_tzformat(self, freq):
             "2013-01-01T00:00:00",
             "2016-01-01T23:59:59",
             freq=freq,
-            tz=pytz.FixedOffset(540),
+            tz=timezone(timedelta(minutes=540)),
         )
         tm.assert_index_equal(idx, expected)
         expected_i8 = date_range(
@@ -665,7 +665,7 @@ def test_constructor_datetime64_tzformat(self, freq):
             "2013-01-01T00:00:00",
             "2016-01-01T23:59:59",
             freq=freq,
-            tz=pytz.FixedOffset(-300),
+            tz=timezone(timedelta(minutes=-300)),
         )
         tm.assert_index_equal(idx, expected)
         # Unable to use `US/Eastern` because of DST
@@ -679,7 +679,7 @@ def test_constructor_datetime64_tzformat(self, freq):
             "2013-01-01T00:00:00",
             "2016-01-01T23:59:59",
             freq=freq,
-            tz=pytz.FixedOffset(540),
+            tz=timezone(timedelta(minutes=540)),
         )
         tm.assert_index_equal(idx, expected)
         expected_i8 = date_range(

pandas/tests/io/parser/test_parse_dates.py

@@ -6,6 +6,8 @@
 from datetime import (
     date,
     datetime,
+    timedelta,
+    timezone,
 )
 from io import StringIO
 import warnings
@@ -935,7 +937,11 @@ def test_parse_tz_aware(all_parsers, request):
         {"x": [0.5]}, index=Index([Timestamp("2012-06-13 01:39:00+00:00")], name="Date")
     )
     tm.assert_frame_equal(result, expected)
-    assert result.index.tz is pytz.utc
+    if parser.engine == "pyarrow":
+        expected_tz = pytz.utc
+    else:
+        expected_tz = timezone.utc
+    assert result.index.tz is expected_tz
 
 
 @xfail_pyarrow
@@ -1563,7 +1569,7 @@ def test_parse_timezone(all_parsers):
                 start="2018-01-04 09:01:00",
                 end="2018-01-04 09:05:00",
                 freq="1min",
-                tz=pytz.FixedOffset(540),
+                tz=timezone(timedelta(minutes=540)),
             )
         ),
         freq=None,

pandas/tests/scalar/timestamp/test_rendering.py

@@ -41,10 +41,8 @@ def test_repr_utcoffset(self):
         date_with_utc_offset = Timestamp("2014-03-13 00:00:00-0400", tz=None)
         assert "2014-03-13 00:00:00-0400" in repr(date_with_utc_offset)
         assert "tzoffset" not in repr(date_with_utc_offset)
-        assert "pytz.FixedOffset(-240)" in repr(date_with_utc_offset)
-        expr = repr(date_with_utc_offset).replace(
-            "'pytz.FixedOffset(-240)'", "pytz.FixedOffset(-240)"
-        )
+        assert "UTC-04:00" in repr(date_with_utc_offset)
+        expr = repr(date_with_utc_offset)
         assert date_with_utc_offset == eval(expr)
 
     def test_timestamp_repr_pre1900(self):

pandas/tests/scalar/timestamp/test_timestamp.py

@@ -4,6 +4,7 @@
 from datetime import (
     datetime,
     timedelta,
+    timezone,
 )
 import locale
 import unicodedata
@@ -12,10 +13,7 @@
 import numpy as np
 import pytest
 import pytz
-from pytz import (
-    timezone,
-    utc,
-)
+from pytz import utc
 
 from pandas._libs.tslibs.dtypes import NpyDatetimeUnit
 from pandas._libs.tslibs.timezones import (
@@ -233,7 +231,7 @@ def test_tz(self):
         assert conv.hour == 19
 
     def test_utc_z_designator(self):
-        assert get_timezone(Timestamp("2014-11-02 01:00Z").tzinfo) is utc
+        assert get_timezone(Timestamp("2014-11-02 01:00Z").tzinfo) is timezone.utc
 
     def test_asm8(self):
         np.random.seed(7_960_929)
@@ -251,7 +249,7 @@ def compare(x, y):
             assert int((Timestamp(x).value - Timestamp(y).value) / 1e9) == 0
 
         compare(Timestamp.now(), datetime.now())
-        compare(Timestamp.now("UTC"), datetime.now(timezone("UTC")))
+        compare(Timestamp.now("UTC"), datetime.now(pytz.timezone("UTC")))
         compare(Timestamp.utcnow(), datetime.utcnow())
         compare(Timestamp.today(), datetime.today())
         current_time = calendar.timegm(datetime.now().utctimetuple())