List indexer on PeriodIndex doesn't coerce strings (#11278)

Author: proost

doc/source/whatsnew/v1.1.0.rst

@@ -148,6 +148,7 @@ Indexing
 ^^^^^^^^
 - Bug in slicing on a :class:`DatetimeIndex` with a partial-timestamp dropping high-resolution indices near the end of a year, quarter, or month (:issue:`31064`)
 - Bug in :meth:`PeriodIndex.get_loc` treating higher-resolution strings differently from :meth:`PeriodIndex.get_value` (:issue:`31172`)
+- Indexer which is list of string has a datetime format fails on :class:`PeriodIndex` or :class:`DatetimeIndex` (:issue:`11278`)
 -
 
 Missing

pandas/core/common.py

@@ -20,13 +20,17 @@
 from pandas.core.dtypes.common import (
     is_array_like,
     is_bool_dtype,
+    is_datetime64_any_dtype,
     is_extension_array_dtype,
     is_integer,
+    is_period_dtype,
 )
 from pandas.core.dtypes.generic import ABCIndex, ABCIndexClass, ABCSeries
 from pandas.core.dtypes.inference import _iterable_not_string
 from pandas.core.dtypes.missing import isna, isnull, notnull  # noqa
 
+from pandas.core.construction import array
+
 
 class SettingWithCopyError(ValueError):
     pass
@@ -231,6 +235,11 @@ def asarray_tuplesafe(values, dtype=None):
     if isinstance(values, list) and dtype in [np.object_, object]:
         return construct_1d_object_array_from_listlike(values)
 
+    if isinstance(values, list) and (
+        is_datetime64_any_dtype(dtype) or is_period_dtype(dtype)
+    ):
+        return array(values, dtype)
+
     result = np.asarray(values, dtype=dtype)
 
     if issubclass(result.dtype.type, str):

pandas/core/indexes/base.py

@@ -3264,7 +3264,8 @@ def _convert_listlike_indexer(self, keyarr, kind=None):
 
     @Appender(_index_shared_docs["_convert_arr_indexer"])
     def _convert_arr_indexer(self, keyarr):
-        keyarr = com.asarray_tuplesafe(keyarr)
+        dtype = self.dtype if isinstance(keyarr, list) else None
+        keyarr = com.asarray_tuplesafe(keyarr, dtype)
         return keyarr
 
     _index_shared_docs[

pandas/core/indexes/datetimes.py

@@ -901,6 +901,9 @@ def indexer_between_time(
 
         return mask.nonzero()[0]
 
+    def _convert_list_indexer(self, keyarr, kind=None):
+        return super().get_indexer_for(keyarr)
+
 
 DatetimeIndex._add_numeric_methods_disabled()
 DatetimeIndex._add_logical_methods_disabled()

pandas/core/indexes/period.py

@@ -823,6 +823,9 @@ def memory_usage(self, deep=False):
             result += self._int64index.memory_usage(deep=deep)
         return result
 
+    def _convert_list_indexer(self, keyarr, kind=None):
+        return super().get_indexer_for(keyarr)
+
 
 PeriodIndex._add_numeric_methods_disabled()
 PeriodIndex._add_logical_methods_disabled()

pandas/tests/indexing/test_loc.py

@@ -6,7 +6,7 @@
 import pytest
 
 import pandas as pd
-from pandas import DataFrame, Series, Timestamp, date_range
+from pandas import DataFrame, Period, Series, Timestamp, date_range, period_range
 import pandas._testing as tm
 from pandas.api.types import is_scalar
 from pandas.tests.indexing.common import Base
@@ -942,6 +942,71 @@ def test_loc_reverse_assignment(self):
 
         tm.assert_series_equal(result, expected)
 
+    @pytest.mark.parametrize(
+        "idx,labels,expected_idx",
+        [
+            (
+                period_range(start="2000", periods=20, freq="D"),
+                ["2000-01-04", "2000-01-08", "2000-01-12"],
+                [
+                    Period("2000-01-04", freq="D"),
+                    Period("2000-01-08", freq="D"),
+                    Period("2000-01-12", freq="D"),
+                ],
+            ),
+            (
+                date_range(start="2000", periods=20, freq="D"),
+                ["2000-01-04", "2000-01-08", "2000-01-12"],
+                [
+                    Timestamp("2000-01-04", freq="D"),
+                    Timestamp("2000-01-08", freq="D"),
+                    Timestamp("2000-01-12", freq="D"),
+                ],
+            ),
+        ],
+    )
+    def test_loc_with_datetime_string_list(self, idx, labels, expected_idx):
+        # GH 11278
+        s = Series(range(20), index=idx)
+        df = DataFrame(range(20), index=idx)
+
+        expected_value = [3, 7, 11]
+        expected_s = Series(expected_value, expected_idx)
+        expected_df = DataFrame(expected_value, expected_idx)
+
+        tm.assert_series_equal(expected_s, s.loc[labels])
+        tm.assert_series_equal(expected_s, s[labels])
+        tm.assert_frame_equal(expected_df, df.loc[labels])
+
+    @pytest.mark.parametrize(
+        "idx,labels,msg",
+        [
+            (
+                period_range(start="2000", periods=20, freq="D"),
+                ["2000-01-04", "2000-01-30"],
+                r"None of \[PeriodIndex\(\['2000-01-04', '2000-01-30'\], "
+                r"dtype='period\[D\]', freq='D'\)\] are in the \[index\]",
+            ),
+            (
+                date_range(start="2000", periods=20, freq="D"),
+                ["2000-01-04", "2000-01-30"],
+                r"None of \[DatetimeIndex\(\['2000-01-04', '2000-01-30'\], "
+                r"dtype='datetime64\[ns\]', freq=None\)\] are in the \[index\]",
+            ),
+        ],
+    )
+    def test_loc_with_datetime_string_and_missing_value(self, idx, labels, msg):
+        # GH 11278
+        s = Series(range(20), index=idx)
+        df = DataFrame(range(20), index=idx)
+
+        with pytest.raises(KeyError, match=msg):
+            s.loc[labels]
+        with pytest.raises(KeyError, match=msg):
+            s[labels]
+        with pytest.raises(KeyError, match=msg):
+            df.loc[labels]
+
 
 def test_series_loc_getitem_label_list_missing_values():
     # gh-11428