TST: More explicit test patterns (#54463)

Author: mroeschke

pandas/tests/arithmetic/test_datetime64.py

@@ -414,8 +414,7 @@ def test_dti_cmp_datetimelike(self, other, tz_naive_fixture):
         dti = date_range("2016-01-01", periods=2, tz=tz)
         if tz is not None:
             if isinstance(other, np.datetime64):
-                # no tzaware version available
-                return
+                pytest.skip("no tzaware version available")
             other = localize_pydatetime(other, dti.tzinfo)
 
         result = dti == other
@@ -1080,13 +1079,14 @@ def test_dt64arr_add_dtlike_raises(self, tz_naive_fixture, box_with_array):
     @pytest.mark.parametrize("freq", ["H", "D", "W", "M", "MS", "Q", "B", None])
     @pytest.mark.parametrize("dtype", [None, "uint8"])
     def test_dt64arr_addsub_intlike(
-        self, dtype, box_with_array, freq, tz_naive_fixture
+        self, request, dtype, box_with_array, freq, tz_naive_fixture
     ):
         # GH#19959, GH#19123, GH#19012
         tz = tz_naive_fixture
         if box_with_array is pd.DataFrame:
-            # alignment headaches
-            return
+            request.node.add_marker(
+                pytest.mark.xfail(raises=ValueError, reason="Axis alignment fails")
+            )
 
         if freq is None:
             dti = DatetimeIndex(["NaT", "2017-04-05 06:07:08"], tz=tz)

pandas/tests/arithmetic/test_numeric.py

@@ -377,7 +377,7 @@ def test_divmod_zero(self, zero, numeric_idx):
     def test_div_negative_zero(self, zero, numeric_idx, op):
         # Check that -1 / -0.0 returns np.inf, not -np.inf
         if numeric_idx.dtype == np.uint64:
-            return
+            pytest.skip(f"Not relevant for {numeric_idx.dtype}")
         idx = numeric_idx - 3
 
         expected = Index([-np.inf, -np.inf, -np.inf, np.nan, np.inf], dtype=np.float64)

pandas/tests/arithmetic/test_timedelta64.py

@@ -1002,7 +1002,7 @@ def test_td64arr_add_sub_datetimelike_scalar(
             ts = dt_scalar.to_pydatetime()
         elif cls is np.datetime64:
             if tz_naive_fixture is not None:
-                return
+                pytest.skip(f"{cls} doesn support {tz_naive_fixture}")
             ts = dt_scalar.to_datetime64()
         else:
             ts = dt_scalar

pandas/tests/frame/methods/test_to_csv.py

@@ -674,10 +674,7 @@ def create_cols(name):
 
         # add in some nans
         df_float.iloc[30:50, 1:3] = np.nan
-
-        # ## this is a bug in read_csv right now ####
-        # df_dt.loc[30:50,1:3] = np.nan
-        # FIXME: don't leave commented-out
+        df_dt.iloc[30:50, 1:3] = np.nan
 
         df = pd.concat([df_float, df_int, df_bool, df_object, df_dt], axis=1)
 

pandas/tests/frame/test_arithmetic.py

@@ -1766,7 +1766,12 @@ def test_inplace_ops_identity(self):
         [
             "add",
             "and",
-            "div",
+            pytest.param(
+                "div",
+                marks=pytest.mark.xfail(
+                    raises=AttributeError, reason="__idiv__ not implemented"
+                ),
+            ),
             "floordiv",
             "mod",
             "mul",
@@ -1778,9 +1783,6 @@ def test_inplace_ops_identity(self):
         ],
     )
     def test_inplace_ops_identity2(self, op):
-        if op == "div":
-            return
-
         df = DataFrame({"a": [1.0, 2.0, 3.0], "b": [1, 2, 3]})
 
         operand = 2

pandas/tests/frame/test_constructors.py

@@ -2514,7 +2514,7 @@ def test_dict_nocopy(
         b = np.array([3, 4], dtype=any_numpy_dtype)
         if b.dtype.kind in ["S", "U"]:
             # These get cast, making the checks below more cumbersome
-            return
+            pytest.skip(f"{b.dtype} get cast, making the checks below more cumbersome")
 
         c = pd.array([1, 2], dtype=any_numeric_ea_dtype)
         c_orig = c.copy()

pandas/tests/groupby/test_api.py

@@ -205,8 +205,7 @@ def test_frame_consistency(groupby_func):
 def test_series_consistency(request, groupby_func):
     # GH#48028
     if groupby_func in ("first", "last"):
-        # first and last are entirely different between Series and groupby
-        return
+        pytest.skip("first and last are entirely different between Series and groupby")
 
     if groupby_func in ("cumcount", "corrwith", "ngroup"):
         assert not hasattr(Series, groupby_func)

pandas/tests/groupby/transform/test_transform.py

@@ -775,9 +775,8 @@ def frame_mi(frame):
         {"by": np.random.default_rng(2).integers(0, 50, size=10).astype(float)},
         {"level": 0},
         {"by": "string"},
-        # {"by": 'string_missing'}]:
-        # {"by": ['int','string']}]:
-        # TODO: remove or enable commented-out code
+        pytest.param({"by": "string_missing"}, marks=pytest.mark.xfail),
+        {"by": ["int", "string"]},
     ],
 )
 def test_cython_transform_frame(request, op, args, targop, df_fix, gb_target):
@@ -825,9 +824,9 @@ def test_cython_transform_frame(request, op, args, targop, df_fix, gb_target):
         {"by": np.random.default_rng(2).integers(0, 50, size=10).astype(float)},
         {"level": 0},
         {"by": "string"},
-        # {"by": 'string_missing'}]:
-        # {"by": ['int','string']}]:
-        # TODO: remove or enable commented-out code
+        # TODO: create xfail condition given other params
+        # {"by": 'string_missing'},
+        {"by": ["int", "string"]},
     ],
 )
 @pytest.mark.parametrize(

pandas/tests/indexes/multi/test_setops.py

@@ -626,8 +626,7 @@ def test_union_with_duplicates_keep_ea_dtype(dupe_val, any_numeric_ea_dtype):
 def test_union_duplicates(index, request):
     # GH#38977
     if index.empty or isinstance(index, (IntervalIndex, CategoricalIndex)):
-        # No duplicates in empty indexes
-        return
+        pytest.skip(f"No duplicates in an empty {type(index).__name__}")
 
     values = index.unique().values.tolist()
     mi1 = MultiIndex.from_arrays([values, [1] * len(values)])

pandas/tests/indexes/test_any_index.py

@@ -34,7 +34,7 @@ def test_hash_error(index):
 
 def test_mutability(index):
     if not len(index):
-        return
+        pytest.skip("Test doesn't make sense for empty index")
     msg = "Index does not support mutable operations"
     with pytest.raises(TypeError, match=msg):
         index[0] = index[0]

pandas/tests/indexes/test_base.py

@@ -561,11 +561,9 @@ def test_map_dictlike_simple(self, mapper):
     def test_map_dictlike(self, index, mapper, request):
         # GH 12756
         if isinstance(index, CategoricalIndex):
-            # Tested in test_categorical
-            return
+            pytest.skip("Tested in test_categorical")
         elif not index.is_unique:
-            # Cannot map duplicated index
-            return
+            pytest.skip("Cannot map duplicated index")
 
         rng = np.arange(len(index), 0, -1, dtype=np.int64)
 
@@ -785,7 +783,7 @@ def test_drop_tuple(self, values, to_drop):
     def test_drop_with_duplicates_in_index(self, index):
         # GH38051
         if len(index) == 0 or isinstance(index, MultiIndex):
-            return
+            pytest.skip("Test doesn't make sense for empty MultiIndex")
         if isinstance(index, IntervalIndex) and not IS64:
             pytest.skip("Cannot test IntervalIndex with int64 dtype on 32 bit platform")
         index = index.unique().repeat(2)

pandas/tests/indexes/test_common.py

@@ -121,10 +121,6 @@ def test_set_name_methods(self, index_flat):
         assert res is None
         assert index.name == new_name
         assert index.names == [new_name]
-        # FIXME: dont leave commented-out
-        # with pytest.raises(TypeError, match="list-like"):
-        #    # should still fail even if it would be the right length
-        #    ind.set_names("a")
         with pytest.raises(ValueError, match="Level must be None"):
             index.set_names("a", level=0)
 
@@ -134,6 +130,12 @@ def test_set_name_methods(self, index_flat):
         assert index.name == name
         assert index.names == [name]
 
+    @pytest.mark.xfail
+    def test_set_names_single_label_no_level(self, index_flat):
+        with pytest.raises(TypeError, match="list-like"):
+            # should still fail even if it would be the right length
+            index_flat.set_names("a")
+
     def test_copy_and_deepcopy(self, index_flat):
         index = index_flat
 

pandas/tests/indexes/test_index_new.py

@@ -132,7 +132,10 @@ def test_constructor_infer_nat_dt_like(
     ):
         if isinstance(nulls_fixture, Decimal):
             # We dont cast these to datetime64/timedelta64
-            return
+            pytest.skip(
+                f"We don't cast {type(nulls_fixture).__name__} to "
+                "datetime64/timedelta64"
+            )
 
         expected = klass([NaT, NaT])
         assert expected.dtype == dtype

pandas/tests/indexes/test_indexing.py

@@ -56,8 +56,7 @@ def test_take_invalid_kwargs(self, index):
     def test_take(self, index):
         indexer = [4, 3, 0, 2]
         if len(index) < 5:
-            # not enough elements; ignore
-            return
+            pytest.skip("Test doesn't make sense since not enough elements")
 
         result = index.take(indexer)
         expected = index[indexer]
@@ -81,7 +80,7 @@ def test_take_minus1_without_fill(self, index):
         # -1 does not get treated as NA unless allow_fill=True is passed
         if len(index) == 0:
             # Test is not applicable
-            return
+            pytest.skip("Test doesn't make sense for empty index")
 
         result = index.take([0, 0, -1])
 
@@ -289,7 +288,7 @@ class TestPutmask:
     def test_putmask_with_wrong_mask(self, index):
         # GH#18368
         if not len(index):
-            return
+            pytest.skip("Test doesn't make sense for empty index")
 
         fill = index[0]
 

pandas/tests/indexes/test_numpy_compat.py

@@ -154,7 +154,7 @@ def test_numpy_ufuncs_other(index, func):
 def test_numpy_ufuncs_reductions(index, func, request):
     # TODO: overlap with tests.series.test_ufunc.test_reductions
     if len(index) == 0:
-        return
+        pytest.skip("Test doesn't make sense for empty index.")
 
     if isinstance(index, CategoricalIndex) and index.dtype.ordered is False:
         with pytest.raises(TypeError, match="is not ordered for"):

pandas/tests/indexes/test_old_base.py

@@ -167,7 +167,7 @@ def test_numeric_compat(self, simple_index):
         #  we can remove multi.test_compat.test_numeric_compat
         assert not isinstance(idx, MultiIndex)
         if type(idx) is Index:
-            return
+            pytest.skip("Not applicable for Index")
         if is_numeric_dtype(simple_index.dtype) or isinstance(
             simple_index, TimedeltaIndex
         ):
@@ -246,9 +246,10 @@ def test_ensure_copied_data(self, index):
             # Needs "freq" specification:
             init_kwargs["freq"] = index.freq
         elif isinstance(index, (RangeIndex, MultiIndex, CategoricalIndex)):
-            # RangeIndex cannot be initialized from data
-            # MultiIndex and CategoricalIndex are tested separately
-            return
+            pytest.skip(
+                "RangeIndex cannot be initialized from data, "
+                "MultiIndex and CategoricalIndex are tested separately"
+            )
         elif index.dtype == object and index.inferred_type == "boolean":
             init_kwargs["dtype"] = index.dtype
 
@@ -319,9 +320,8 @@ def test_memory_usage(self, index):
             assert result3 > result2
 
     def test_argsort(self, index):
-        # separately tested
         if isinstance(index, CategoricalIndex):
-            return
+            pytest.skip(f"{type(self).__name__} separately tested")
 
         result = index.argsort()
         expected = np.array(index).argsort()
@@ -400,7 +400,7 @@ def test_insert_base(self, index):
         result = index[1:4]
 
         if not len(index):
-            return
+            pytest.skip("Not applicable for empty index")
 
         # test 0th element
         assert index[0:4].equals(result.insert(0, index[0]))
@@ -434,11 +434,11 @@ def test_insert_out_of_bounds(self, index):
 
     def test_delete_base(self, index):
         if not len(index):
-            return
+            pytest.skip("Not applicable for empty index")
 
         if isinstance(index, RangeIndex):
             # tested in class
-            return
+            pytest.skip(f"{type(self).__name__} tested elsewhere")
 
         expected = index[1:]
         result = index.delete(0)
@@ -458,9 +458,7 @@ def test_delete_base(self, index):
     @pytest.mark.filterwarnings(r"ignore:PeriodDtype\[B\] is deprecated:FutureWarning")
     def test_equals(self, index):
         if isinstance(index, IntervalIndex):
-            # IntervalIndex tested separately, the index.equals(index.astype(object))
-            #  fails for IntervalIndex
-            return
+            pytest.skip(f"{type(index).__name__} tested elsewhere")
 
         is_ea_idx = type(index) is Index and not isinstance(index.dtype, np.dtype)
 
@@ -566,12 +564,11 @@ def test_format_empty(self, simple_index):
     def test_fillna(self, index):
         # GH 11343
         if len(index) == 0:
-            return
+            pytest.skip("Not relevant for empty index")
         elif index.dtype == bool:
-            # can't hold NAs
-            return
+            pytest.skip(f"{index.dtype} cannot hold NAs")
         elif isinstance(index, Index) and is_integer_dtype(index.dtype):
-            return
+            pytest.skip(f"Not relevant for Index with {index.dtype}")
         elif isinstance(index, MultiIndex):
             idx = index.copy(deep=True)
             msg = "isna is not defined for MultiIndex"
@@ -654,15 +651,9 @@ def test_map(self, simple_index):
         ],
     )
     @pytest.mark.filterwarnings(r"ignore:PeriodDtype\[B\] is deprecated:FutureWarning")
-    def test_map_dictlike(self, mapper, simple_index):
+    def test_map_dictlike(self, mapper, simple_index, request):
         idx = simple_index
-        if isinstance(idx, CategoricalIndex):
-            # FIXME: this fails with CategoricalIndex bc it goes through
-            # Categorical.map which ends up calling get_indexer with
-            #  non-unique values, which raises.  This _should_ work fine for
-            #  CategoricalIndex.
-            pytest.skip(f"skipping tests for {type(idx)}")
-        elif isinstance(idx, (DatetimeIndex, TimedeltaIndex, PeriodIndex)):
+        if isinstance(idx, (DatetimeIndex, TimedeltaIndex, PeriodIndex)):
             pytest.skip("Tested elsewhere.")
 
         identity = mapper(idx.values, idx)

pandas/tests/indexes/test_setops.py

@@ -194,7 +194,7 @@ def test_set_ops_error_cases(self, case, method, index):
     @pytest.mark.filterwarnings(r"ignore:PeriodDtype\[B\] is deprecated:FutureWarning")
     def test_intersection_base(self, index):
         if isinstance(index, CategoricalIndex):
-            return
+            pytest.skip(f"Not relevant for {type(index).__name__}")
 
         first = index[:5]
         second = index[:3]
@@ -280,13 +280,13 @@ def test_difference_base(self, sort, index):
     )
     def test_symmetric_difference(self, index):
         if isinstance(index, CategoricalIndex):
-            return
+            pytest.skip(f"Not relevant for {type(index).__name__}")
         if len(index) < 2:
-            return
+            pytest.skip("Too few values for test")
         if index[0] in index[1:] or index[-1] in index[:-1]:
             # index fixture has e.g. an index of bools that does not satisfy this,
             #  another with [0, 0, 1, 1, 2, 2]
-            return
+            pytest.skip("Index values no not satisfy test condition.")
 
         first = index[1:]
         second = index[:-1]
@@ -465,7 +465,7 @@ def test_difference_preserves_type_empty(self, index, sort):
         # If taking difference of a set and itself, it
         # needs to preserve the type of the index
         if not index.is_unique:
-            return
+            pytest.skip("Not relevant since index is not unique")
         result = index.difference(index, sort=sort)
         expected = index[:0]
         tm.assert_index_equal(result, expected, exact=True)
@@ -488,7 +488,7 @@ def test_intersection_difference_match_empty(self, index, sort):
         # empty index produces the same index as the difference
         # of an index with itself.  Test for all types
         if not index.is_unique:
-            return
+            pytest.skip("Not relevant because index is not unique")
         inter = index.intersection(index[:0])
         diff = index.difference(index, sort=sort)
         tm.assert_index_equal(inter, diff, exact=True)
@@ -521,7 +521,7 @@ def test_intersection_duplicates_all_indexes(index):
     # GH#38743
     if index.empty:
         # No duplicates in empty indexes
-        return
+        pytest.skip("Not relevant for empty Index")
 
     idx = index
     idx_non_unique = idx[[0, 0, 1, 2]]