CLN: Old timezone issues

doc/source/whatsnew/v0.24.0.txt

@@ -561,6 +561,10 @@ Timezones
 - Fixed bug where :meth:`DataFrame.describe` and :meth:`Series.describe` on tz-aware datetimes did not show `first` and `last` result (:issue:`21328`)
 - Bug in :class:`DatetimeIndex` comparisons failing to raise ``TypeError`` when comparing timezone-aware ``DatetimeIndex`` against ``np.datetime64`` (:issue:`22074`)
 - Bug in ``DataFrame`` assignment with a timezone-aware scalar (:issue:`19843`)
+- Bug when constructing a :class:`DatetimeIndex` with :class:`Timestamp`s constructed with the ``replace`` method across DST (:issue:`18785`)
+- Bug when setting a new value with :meth:`DataFrame.loc` with a :class:`DatetimeIndex` with a DST transition (:issue:`18308`, :issue:`20724`)
+- Bug in :meth:`DatetimeIndex.unique` that did not re-localize tz-aware dates correctly (:issue:`21737`)
+- Bug when indexing a :class:`Series` with a DST transition (:issue:`21846`)
 
 Offsets
 ^^^^^^^

pandas/tests/indexes/datetimes/test_construction.py

@@ -503,6 +503,24 @@ def test_construction_int_rountrip(self, tz_naive_fixture):
         expected = DatetimeIndex([1293858000000000000], tz=tz).asi8[0]
         assert result == expected
 
+    def test_construction_from_replaced_timestamps_with_dst(self):
+        # GH 18785
+        index = pd.date_range(pd.Timestamp(2000, 1, 1),
+                              pd.Timestamp(2005, 1, 1),
+                              freq='MS', tz='Australia/Melbourne')
+        test = pd.DataFrame({'data': range(len(index))}, index=index)
+        test = test.resample('Y').mean()
+        result = pd.DatetimeIndex([x.replace(month=6, day=1)
+                                   for x in test.index])
+        expected = pd.DatetimeIndex(['2000-06-01 00:00:00',
+                                     '2001-06-01 00:00:00',
+                                     '2002-06-01 00:00:00',
+                                     '2003-06-01 00:00:00',
+                                     '2004-06-01 00:00:00',
+                                     '2005-06-01 00:00:00'],
+                                    tz='Australia/Melbourne')
+        tm.assert_index_equal(result, expected)
+
 
 class TestTimeSeries(object):
 

pandas/tests/indexes/datetimes/test_datetime.py

@@ -394,3 +394,6 @@ def test_factorize_dst(self):
     def test_unique(self, arr, expected):
         result = arr.unique()
         tm.assert_index_equal(result, expected)
+        # GH 21737
+        # Ensure the underlying data is consistent
+        assert result[0] == expected[0]

pandas/tests/indexing/test_datetime.py

@@ -1,5 +1,9 @@
+from datetime import datetime, timedelta
+
 import numpy as np
 import pandas as pd
+from dateutil import tz
+
 from pandas import date_range, Index, DataFrame, Series, Timestamp
 from pandas.util import testing as tm
 
@@ -266,3 +270,46 @@ def test_nanosecond_getitem_setitem_with_tz(self):
         result.loc[df.index[0], 'a'] = -1
         expected = DataFrame(-1, index=index, columns=['a'])
         tm.assert_frame_equal(result, expected)
+
+    def test_loc_getitem_across_dst(self):
+        # GH 21846
+        idx = pd.date_range('2017-10-29 01:30:00',
+                            tz='Europe/Berlin', periods=5, freq='30 min')
+        series2 = pd.Series([0, 1, 2, 3, 4],
+                            index=idx)
+
+        t_1 = pd.Timestamp('2017-10-29 02:30:00+02:00', tz='Europe/Berlin',
+                           freq='30min')
+        t_2 = pd.Timestamp('2017-10-29 02:00:00+01:00', tz='Europe/Berlin',
+                           freq='30min')
+        result = series2.loc[t_1:t_2]
+        expected = pd.Series([2, 3], index=idx[2:4])
+        tm.assert_series_equal(result, expected)
+
+        result = series2[t_1]
+        expected = 2
+        assert result == expected
+
+    def test_loc_incremental_setitem_with_dst(self):
+        # GH 20724
+        base = datetime(2015, 11, 1, tzinfo=tz.gettz("US/Pacific"))
+        idxs = [base + timedelta(seconds=i * 900) for i in range(16)]
+        result = pd.Series([0], index=[idxs[0]])
+        for ts in idxs:
+            result.loc[ts] = 1
+        expected = pd.Series(1, index=idxs)
+        tm.assert_series_equal(result, expected)
+
+    def test_loc_setitem_with_existing_dst(self):
+        # GH 18308
+        start = pd.Timestamp('2017-10-29 00:00:00+0200', tz='Europe/Madrid')
+        end = pd.Timestamp('2017-10-29 03:00:00+0100', tz='Europe/Madrid')
+        ts = pd.Timestamp('2016-10-10 03:00:00', tz='Europe/Madrid')
+        idx = pd.date_range(start, end, closed='left', freq="H")
+        result = pd.DataFrame(index=idx, columns=['value'])
+        result.loc[ts, 'value'] = 12
+        expected = pd.DataFrame([np.nan] * len(idx) + [12],
+                                index=idx.append(pd.DatetimeIndex([ts])),
+                                columns=['value'],
+                                dtype=object)
+        tm.assert_frame_equal(result, expected)