TST: Use more xfail instead of skip (pandas-dev#45719)

Author: mroeschke

pandas/tests/apply/test_series_apply.py

@@ -284,8 +284,6 @@ def test_transform_partial_failure(op, request):
                 raises=AssertionError, reason=f"{op} is successful on any dtype"
             )
         )
-    if op in ("rank", "fillna"):
-        pytest.skip(f"{op} doesn't raise TypeError on object")
 
     # Using object makes most transform kernels fail
     ser = Series(3 * [object])
@@ -497,9 +495,13 @@ def test_map(datetime_series):
     tm.assert_series_equal(a.map(c), exp)
 
 
-def test_map_empty(index):
+def test_map_empty(request, index):
     if isinstance(index, MultiIndex):
-        pytest.skip("Initializing a Series from a MultiIndex is not supported")
+        request.node.add_marker(
+            pytest.mark.xfail(
+                reason="Initializing a Series from a MultiIndex is not supported"
+            )
+        )
 
     s = Series(index)
     result = s.map({})

pandas/tests/base/test_misc.py

@@ -148,14 +148,18 @@ def test_memory_usage_components_narrow_series(dtype):
     assert total_usage == non_index_usage + index_usage
 
 
-def test_searchsorted(index_or_series_obj):
+def test_searchsorted(request, index_or_series_obj):
     # numpy.searchsorted calls obj.searchsorted under the hood.
     # See gh-12238
     obj = index_or_series_obj
 
     if isinstance(obj, pd.MultiIndex):
         # See gh-14833
-        pytest.skip("np.searchsorted doesn't work on pd.MultiIndex")
+        request.node.add_marker(
+            pytest.mark.xfail(
+                reason="np.searchsorted doesn't work on pd.MultiIndex: GH 14833"
+            )
+        )
 
     max_obj = max(obj, default=0)
     index = np.searchsorted(obj, max_obj)

pandas/tests/extension/test_sparse.py

@@ -453,7 +453,7 @@ def _skip_if_different_combine(self, data):
             # arith ops call on dtype.fill_value so that the sparsity
             # is maintained. Combine can't be called on a dtype in
             # general, so we can't make the expected. This is tested elsewhere
-            raise pytest.skip("Incorrected expected from Series.combine")
+            pytest.skip("Incorrected expected from Series.combine and tested elsewhere")
 
     def test_arith_series_with_scalar(self, data, all_arithmetic_operators):
         self._skip_if_different_combine(data)

pandas/tests/frame/test_reductions.py

@@ -1471,9 +1471,15 @@ def test_reductions_deprecation_level_argument(
         with tm.assert_produces_warning(FutureWarning, match="level"):
             getattr(obj, reduction_functions)(level=0)
 
-    def test_reductions_skipna_none_raises(self, frame_or_series, reduction_functions):
-        if reduction_functions in ["count", "mad"]:
-            pytest.skip("Count does not accept skipna. Mad needs a deprecation cycle.")
+    def test_reductions_skipna_none_raises(
+        self, request, frame_or_series, reduction_functions
+    ):
+        if reduction_functions == "count":
+            request.node.add_marker(
+                pytest.mark.xfail(reason="Count does not accept skipna")
+            )
+        elif reduction_functions == "mad":
+            pytest.skip("Mad needs a deprecation cycle: GH 11787")
         obj = frame_or_series([1, 2, 3])
         msg = 'For argument "skipna" expected type bool, received type NoneType.'
         with pytest.raises(ValueError, match=msg):

pandas/tests/groupby/test_groupby.py

@@ -713,11 +713,8 @@ def test_ops_not_as_index(reduction_func):
     # GH 10355, 21090
     # Using as_index=False should not modify grouped column
 
-    if reduction_func in ("corrwith",):
-        pytest.skip("Test not applicable")
-
-    if reduction_func in ("nth", "ngroup"):
-        pytest.skip("Skip until behavior is determined (GH #5755)")
+    if reduction_func in ("corrwith", "nth", "ngroup"):
+        pytest.skip(f"GH 5755: Test not applicable for {reduction_func}")
 
     df = DataFrame(np.random.randint(0, 5, size=(100, 2)), columns=["a", "b"])
     expected = getattr(df.groupby("a"), reduction_func)()
@@ -2268,7 +2265,7 @@ def test_groupby_duplicate_index():
 @pytest.mark.filterwarnings("ignore:tshift is deprecated:FutureWarning")
 def test_dup_labels_output_shape(groupby_func, idx):
     if groupby_func in {"size", "ngroup", "cumcount"}:
-        pytest.skip("Not applicable")
+        pytest.skip(f"Not applicable for {groupby_func}")
     # TODO(2.0) Remove after pad/backfill deprecation enforced
     groupby_func = maybe_normalize_deprecated_kernels(groupby_func)
     df = DataFrame([[1, 1]], columns=idx)

pandas/tests/groupby/test_groupby_subclass.py

@@ -24,7 +24,7 @@ def test_groupby_preserves_subclass(obj, groupby_func):
     # GH28330 -- preserve subclass through groupby operations
 
     if isinstance(obj, Series) and groupby_func in {"corrwith"}:
-        pytest.skip("Not applicable")
+        pytest.skip(f"Not applicable for Series and {groupby_func}")
     # TODO(2.0) Remove after pad/backfill deprecation enforced
     groupby_func = maybe_normalize_deprecated_kernels(groupby_func)
     grouped = obj.groupby(np.arange(0, 10))

pandas/tests/io/excel/test_readers.py

@@ -1562,7 +1562,7 @@ def test_corrupt_files_closed(self, engine, read_ext):
         # GH41778
         errors = (BadZipFile,)
         if engine is None:
-            pytest.skip()
+            pytest.skip(f"Invalid test for engine={engine}")
         elif engine == "xlrd":
             import xlrd
 

pandas/tests/io/json/test_pandas.py

@@ -1553,9 +1553,20 @@ def test_timedelta_as_label(self, date_format, key):
             ("index", "{\"('a', 'b')\":{\"('c', 'd')\":1}}"),
             ("columns", "{\"('c', 'd')\":{\"('a', 'b')\":1}}"),
             # TODO: the below have separate encoding procedures
-            # They produce JSON but not in a consistent manner
-            pytest.param("split", "", marks=pytest.mark.skip),
-            pytest.param("table", "", marks=pytest.mark.skip),
+            pytest.param(
+                "split",
+                "",
+                marks=pytest.mark.xfail(
+                    reason="Produces JSON but not in a consistent manner"
+                ),
+            ),
+            pytest.param(
+                "table",
+                "",
+                marks=pytest.mark.xfail(
+                    reason="Produces JSON but not in a consistent manner"
+                ),
+            ),
         ],
     )
     def test_tuple_labels(self, orient, expected):

pandas/tests/io/json/test_ujson.py

@@ -997,9 +997,11 @@ def test_dataframe_nested(self, orient):
         }
         assert ujson.decode(ujson.encode(nested, **kwargs)) == exp
 
-    def test_dataframe_numpy_labelled(self, orient):
+    def test_dataframe_numpy_labelled(self, orient, request):
         if orient in ("split", "values"):
-            pytest.skip("Incompatible with labelled=True")
+            request.node.add_marker(
+                pytest.mark.xfail(reason=f"{orient} incompatible for labelled=True")
+            )
 
         df = DataFrame(
             [[1, 2, 3], [4, 5, 6]],

pandas/tests/io/parser/dtypes/test_dtypes_basic.py

@@ -193,7 +193,7 @@ def test_delimiter_with_usecols_and_parse_dates(all_parsers):
 @pytest.mark.parametrize("thousands", ["_", None])
 def test_decimal_and_exponential(python_parser_only, numeric_decimal, thousands):
     # GH#31920
-    decimal_number_check(python_parser_only, numeric_decimal, thousands)
+    decimal_number_check(python_parser_only, numeric_decimal, thousands, None)
 
 
 @pytest.mark.parametrize("thousands", ["_", None])
@@ -203,21 +203,22 @@ def test_1000_sep_decimal_float_precision(
 ):
     # test decimal and thousand sep handling in across 'float_precision'
     # parsers
-    decimal_number_check(c_parser_only, numeric_decimal, thousands)
+    decimal_number_check(c_parser_only, numeric_decimal, thousands, float_precision)
     text, value = numeric_decimal
     text = " " + text + " "
     if isinstance(value, str):  # the negative cases (parse as text)
         value = " " + value + " "
-    decimal_number_check(c_parser_only, (text, value), thousands)
+    decimal_number_check(c_parser_only, (text, value), thousands, float_precision)
 
 
-def decimal_number_check(parser, numeric_decimal, thousands):
+def decimal_number_check(parser, numeric_decimal, thousands, float_precision):
     # GH#31920
     value = numeric_decimal[0]
     if thousands is None and "_" in value:
         pytest.skip("Skip test if no thousands sep is defined and sep is in value")
     df = parser.read_csv(
         StringIO(value),
+        float_precision=float_precision,
         sep="|",
         thousands=thousands,
         decimal=",",

pandas/tests/io/parser/test_parse_dates.py

@@ -1703,11 +1703,15 @@ def _helper_hypothesis_delimited_date(call, date_string, **kwargs):
     "date_format",
     ["%d %m %Y", "%m %d %Y", "%m %Y", "%Y %m %d", "%y %m %d", "%Y%m%d", "%y%m%d"],
 )
-def test_hypothesis_delimited_date(date_format, dayfirst, delimiter, test_datetime):
+def test_hypothesis_delimited_date(
+    request, date_format, dayfirst, delimiter, test_datetime
+):
     if date_format == "%m %Y" and delimiter == ".":
-        pytest.skip(
-            "parse_datetime_string cannot reliably tell whether "
-            "e.g. %m.%Y is a float or a date, thus we skip it"
+        request.node.add_marker(
+            pytest.mark.xfail(
+                reason="parse_datetime_string cannot reliably tell whether "
+                "e.g. %m.%Y is a float or a date"
+            )
         )
     result, expected = None, None
     except_in_dateutil, except_out_dateutil = None, None

pandas/tests/io/test_parquet.py

@@ -942,11 +942,17 @@ def test_timestamp_nanoseconds(self, pa):
         df = pd.DataFrame({"a": pd.date_range("2017-01-01", freq="1n", periods=10)})
         check_round_trip(df, pa, write_kwargs={"version": ver})
 
-    def test_timezone_aware_index(self, pa, timezone_aware_date_list):
-        if not pa_version_under2p0:
-            # temporary skip this test until it is properly resolved
-            # https://github.com/pandas-dev/pandas/issues/37286
-            pytest.skip()
+    def test_timezone_aware_index(self, request, pa, timezone_aware_date_list):
+        if (
+            not pa_version_under2p0
+            and timezone_aware_date_list.tzinfo != datetime.timezone.utc
+        ):
+            request.node.add_marker(
+                pytest.mark.xfail(
+                    reason="temporary skip this test until it is properly resolved: "
+                    "https://github.com/pandas-dev/pandas/issues/37286"
+                )
+            )
         idx = 5 * [timezone_aware_date_list]
         df = pd.DataFrame(index=idx, data={"index_as_col": idx})
 
@@ -995,7 +1001,6 @@ def test_basic(self, fp, df_full):
         df["timedelta"] = pd.timedelta_range("1 day", periods=3)
         check_round_trip(df, fp)
 
-    @pytest.mark.skip(reason="not supported")
     def test_duplicate_columns(self, fp):
 
         # not currently able to handle duplicate columns

pandas/tests/resample/test_base.py

@@ -98,11 +98,15 @@ def test_raises_on_non_datetimelike_index():
 
 @all_ts
 @pytest.mark.parametrize("freq", ["M", "D", "H"])
-def test_resample_empty_series(freq, empty_series_dti, resample_method):
+def test_resample_empty_series(freq, empty_series_dti, resample_method, request):
     # GH12771 & GH12868
 
-    if resample_method == "ohlc":
-        pytest.skip("need to test for ohlc from GH13083")
+    if resample_method == "ohlc" and isinstance(empty_series_dti.index, PeriodIndex):
+        request.node.add_marker(
+            pytest.mark.xfail(
+                reason=f"GH13083: {resample_method} fails for PeriodIndex"
+            )
+        )
 
     ser = empty_series_dti
     result = getattr(ser.resample(freq), resample_method)()

pandas/tests/scalar/test_nat.py

@@ -110,9 +110,11 @@ def test_identity(klass, value):
 
 @pytest.mark.parametrize("klass", [Timestamp, Timedelta, Period])
 @pytest.mark.parametrize("value", ["", "nat", "NAT", None, np.nan])
-def test_equality(klass, value):
+def test_equality(klass, value, request):
     if klass is Period and value == "":
-        pytest.skip("Period cannot parse empty string")
+        request.node.add_marker(
+            pytest.mark.xfail(reason="Period cannot parse empty string")
+        )
 
     assert klass(value).value == iNaT
 

pandas/tests/series/methods/test_interpolate.py

@@ -778,7 +778,7 @@ def test_interp_non_timedelta_index(self, interp_methods_ind, ind):
             with pytest.raises(ValueError, match=expected_error):
                 df[0].interpolate(method=method, **kwargs)
 
-    def test_interpolate_timedelta_index(self, interp_methods_ind):
+    def test_interpolate_timedelta_index(self, request, interp_methods_ind):
         """
         Tests for non numerical index types  - object, period, timedelta
         Note that all methods except time, index, nearest and values
@@ -792,14 +792,15 @@ def test_interpolate_timedelta_index(self, interp_methods_ind):
         if method == "pchip":
             pytest.importorskip("scipy")
 
-        if method in {"linear", "pchip"}:
-            result = df[0].interpolate(method=method, **kwargs)
-            expected = Series([0.0, 1.0, 2.0, 3.0], name=0, index=ind)
-            tm.assert_series_equal(result, expected)
-        else:
-            pytest.skip(
-                "This interpolation method is not supported for Timedelta Index yet."
+        if method in {"cubic", "zero"}:
+            request.node.add_marker(
+                pytest.mark.xfail(
+                    reason=f"{method} interpolation is not supported for TimedeltaIndex"
+                )
             )
+        result = df[0].interpolate(method=method, **kwargs)
+        expected = Series([0.0, 1.0, 2.0, 3.0], name=0, index=ind)
+        tm.assert_series_equal(result, expected)
 
     @pytest.mark.parametrize(
         "ascending, expected_values",

pandas/tests/tools/test_to_numeric.py

@@ -390,19 +390,19 @@ def test_timedelta(transform_assert_equal):
     assert_equal(result, expected)
 
 
-def test_period(transform_assert_equal):
+def test_period(request, transform_assert_equal):
     transform, assert_equal = transform_assert_equal
 
     idx = pd.period_range("2011-01", periods=3, freq="M", name="")
     inp = transform(idx)
 
-    if isinstance(inp, Index):
-        result = to_numeric(inp)
-        expected = transform(idx.asi8)
-        assert_equal(result, expected)
-    else:
-        # TODO: PeriodDtype, so support it in to_numeric.
-        pytest.skip("Missing PeriodDtype support in to_numeric")
+    if not isinstance(inp, Index):
+        request.node.add_marker(
+            pytest.mark.xfail(reason="Missing PeriodDtype support in to_numeric")
+        )
+    result = to_numeric(inp)
+    expected = transform(idx.asi8)
+    assert_equal(result, expected)
 
 
 @pytest.mark.parametrize(