more fixes for infer_string mode

Author: jorisvandenbossche

pandas/tests/dtypes/test_concat.py

@@ -47,7 +47,7 @@ def test_concat_periodarray_2d():
         _concat.concat_compat([arr[:2], arr[2:]], axis=1)
 
 
-def test_concat_series_between_empty_and_tzaware_series():
+def test_concat_series_between_empty_and_tzaware_series(using_infer_string):
     tzaware_time = pd.Timestamp("2020-01-01T00:00:00+00:00")
     ser1 = Series(index=[tzaware_time], data=0, dtype=float)
     ser2 = Series(dtype=float)
@@ -57,7 +57,9 @@ def test_concat_series_between_empty_and_tzaware_series():
         data=[
             (0.0, None),
         ],
-        index=pd.Index([tzaware_time], dtype=object),
+        index=[tzaware_time]
+        if using_infer_string
+        else pd.Index([tzaware_time], dtype=object),
         columns=[0, 1],
         dtype=float,
     )

pandas/tests/frame/constructors/test_from_dict.py

@@ -3,8 +3,6 @@
 import numpy as np
 import pytest
 
-from pandas._config import using_string_dtype
-
 from pandas import (
     DataFrame,
     Index,
@@ -44,7 +42,6 @@ def test_constructor_single_row(self):
         )
         tm.assert_frame_equal(result, expected)
 
-    @pytest.mark.xfail(using_string_dtype(), reason="columns inferring logic broken")
     def test_constructor_list_of_series(self):
         data = [
             OrderedDict([["a", 1.5], ["b", 3.0], ["c", 4.0]]),

pandas/tests/frame/test_constructors.py

@@ -21,8 +21,6 @@
 from numpy.ma import mrecords
 import pytest
 
-from pandas._config import using_string_dtype
-
 from pandas._libs import lib
 from pandas.compat.numpy import np_version_gt2
 from pandas.errors import IntCastingNaNError
@@ -1974,7 +1972,6 @@ def test_constructor_with_datetimes4(self):
         df = DataFrame({"value": dr})
         assert str(df.iat[0, 0].tz) == "US/Eastern"
 
-    @pytest.mark.xfail(using_string_dtype(), reason="TODO(infer_string)")
     def test_constructor_with_datetimes5(self):
         # GH 7822
         # preserver an index with a tz on dict construction

pandas/tests/indexes/base_class/test_setops.py

@@ -240,7 +240,6 @@ def test_tuple_union_bug(self, method, expected, sort):
     def test_union_name_preservation(
         self, first_list, second_list, first_name, second_name, expected_name, sort
     ):
-        # expected_dtype = object if not first_list or not second_list else "str"
         first = Index(first_list, name=first_name)
         second = Index(second_list, name=second_name)
         union = first.union(second, sort=sort)

pandas/tests/indexes/datetimes/test_join.py

@@ -70,13 +70,17 @@ def test_join_utc_convert(self, join_type):
         assert isinstance(result, DatetimeIndex)
         assert result.tz is timezone.utc
 
-    def test_datetimeindex_union_join_empty(self, sort):
+    def test_datetimeindex_union_join_empty(self, sort, using_infer_string):
         dti = date_range(start="1/1/2001", end="2/1/2001", freq="D")
         empty = Index([])
 
         result = dti.union(empty, sort=sort)
-        expected = dti.astype("O")
-        tm.assert_index_equal(result, expected)
+        if using_infer_string:
+            assert isinstance(result, DatetimeIndex)
+            tm.assert_index_equal(result, dti)
+        else:
+            expected = dti.astype("O")
+            tm.assert_index_equal(result, expected)
 
         result = dti.join(empty)
         assert isinstance(result, DatetimeIndex)

pandas/tests/indexing/test_loc.py

@@ -2000,7 +2000,6 @@ def test_loc_setitem_datetime_keys_cast(self, conv, using_infer_string):
         df.loc[conv(dt1), "one"] = 100
         df.loc[conv(dt2), "one"] = 200
 
-        # breakpoint()
         expected = DataFrame(
             {"one": [100.0, 200.0]},
             index=Index(

pandas/tests/series/methods/test_combine_first.py

@@ -31,7 +31,7 @@ def test_combine_first_name(self, datetime_series):
         result = datetime_series.combine_first(datetime_series[:5])
         assert result.name == datetime_series.name
 
-    def test_combine_first(self):
+    def test_combine_first(self, using_infer_string):
         values = np.arange(20, dtype=np.float64)
         series = Series(values, index=np.arange(20, dtype=np.int64))
 
@@ -64,7 +64,8 @@ def test_combine_first(self):
         ser = Series([1.0, 2, 3], index=[0, 1, 2])
         empty = Series([], index=[], dtype=object)
         result = ser.combine_first(empty)
-        ser.index = ser.index.astype("O")
+        if not using_infer_string:
+            ser.index = ser.index.astype("O")
         tm.assert_series_equal(result, ser.astype(object))
 
     def test_combine_first_dt64(self, unit):