BUG: make pct_change can handle the anchored freq #28664 (#28681)

Author: dongho-jung

doc/source/whatsnew/v1.0.0.rst

@@ -440,6 +440,7 @@ Reshaping
 - :func:`qcut` and :func:`cut` now handle boolean input (:issue:`20303`)
 - Fix to ensure all int dtypes can be used in :func:`merge_asof` when using a tolerance value. Previously every non-int64 type would raise an erroneous ``MergeError`` (:issue:`28870`).
 - Better error message in :func:`get_dummies` when `columns` isn't a list-like value (:issue:`28383`)
+- Bug :meth:`Series.pct_change` where supplying an anchored frequency would throw a ValueError (:issue:`28664`)
 
 Sparse
 ^^^^^^

pandas/core/generic.py

@@ -10443,6 +10443,7 @@ def pct_change(self, periods=1, fill_method="pad", limit=None, freq=None, **kwar
             data = self.fillna(method=fill_method, limit=limit, axis=axis)
 
         rs = data.div(data.shift(periods=periods, freq=freq, axis=axis, **kwargs)) - 1
+        rs = rs.loc[~rs.index.duplicated()]
         rs = rs.reindex_like(data)
         if freq is None:
             mask = isna(com.values_from_object(data))

pandas/tests/series/test_timeseries.py

@@ -370,6 +370,16 @@ def test_pct_change(self, datetime_series):
             rs, (filled / filled.shift(freq="5D") - 1).reindex_like(filled)
         )
 
+    def test_pct_change_with_duplicate_axis(self):
+        # GH 28664
+        common_idx = date_range("2019-11-14", periods=5, freq="D")
+        result = Series(range(5), common_idx).pct_change(freq="B")
+
+        # the reason that the expected should be like this is documented at PR 28681
+        expected = Series([np.NaN, np.inf, np.NaN, np.NaN, 3.0], common_idx)
+
+        tm.assert_series_equal(result, expected)
+
     def test_pct_change_shift_over_nas(self):
         s = Series([1.0, 1.5, np.nan, 2.5, 3.0])