cleanup

Signed-off-by: Anatoly Myachev <anatoly.myachev@intel.com>

tests/column/shift_test.py

@@ -1,43 +1,25 @@
-import pandas as pd
-import polars as pl
-import pytest
-from polars.testing import assert_frame_equal
-
 from tests.utils import BaseHandler
 from tests.utils import compare_dataframe_with_reference
 from tests.utils import float_dataframe_1
 from tests.utils import integer_dataframe_1
 
 
 def test_shift_with_fill_value(library: BaseHandler) -> None:
-    if library.name == "modin":
-        pytest.skip("TODO: enable for modin")
     df = integer_dataframe_1(library)
     ns = df.__dataframe_namespace__()
     result = df.assign(df.col("a").shift(1).fill_null(999))
-    if library.name == "pandas-numpy":
+    if library.name in ("pandas-numpy", "modin"):
         result = result.cast({name: ns.Int64() for name in ("a", "b")})
     expected = {"a": [999, 1, 2], "b": [4, 5, 6]}
     compare_dataframe_with_reference(result, expected, dtype=ns.Int64)
 
 
 def test_shift_without_fill_value(library: BaseHandler) -> None:
-    if library.name == "modin":
-        pytest.skip("TODO: enable for modin")
     df = float_dataframe_1(library)
+    ns = df.__dataframe_namespace__()
     result = df.assign(df.col("a").shift(-1))
-    if library.name == "pandas-numpy":
-        expected = pd.DataFrame({"a": [3.0, float("nan")]})
-        pd.testing.assert_frame_equal(result.dataframe, expected)
-    elif library.name == "pandas-nullable":
-        expected = pd.DataFrame({"a": [3.0, None]}, dtype="Float64")
-        pd.testing.assert_frame_equal(result.dataframe, expected)
-    elif library.name == "polars-lazy":
-        expected = pl.DataFrame({"a": [3.0, None]})
-        assert_frame_equal(result.dataframe.collect(), expected)  # type: ignore[attr-defined]
-    else:  # pragma: no cover
-        msg = "unexpected library"
-        raise AssertionError(msg)
+    expected = {"a": [3.0, float("nan")]}
+    compare_dataframe_with_reference(result, expected, dtype=ns.Float64)
 
 
 def test_shift_with_fill_value_complicated(library: BaseHandler) -> None:

tests/conftest.py

@@ -3,6 +3,8 @@
 import sys
 from typing import Any
 
+import pytest
+
 from tests.utils import ModinHandler
 from tests.utils import PandasHandler
 from tests.utils import PolarsHandler
@@ -56,3 +58,17 @@ def pytest_generate_tests(metafunc: Any) -> None:
         lib_handlers = [LIBRARIES_HANDLERS[lib] for lib in libraries]
 
         metafunc.parametrize("library", lib_handlers, ids=libraries)
+
+
+ci_xfail_ids = [
+    # https://github.com/modin-project/modin/issues/7212
+    "join_test.py::test_join_left[modin]",
+    "join_test.py::test_join_two_keys[modin]",
+    "persistedness_test.py::test_cross_df_propagation[modin]",
+]
+
+
+def pytest_collection_modifyitems(items: list[pytest.Item]) -> None:
+    for item in items:
+        if any(id_ in item.nodeid for id_ in ci_xfail_ids):
+            item.add_marker(pytest.mark.xfail(strict=True))

tests/dataframe/join_test.py

@@ -11,8 +11,6 @@
 
 
 def test_join_left(library: BaseHandler) -> None:
-    if library.name == "modin":
-        pytest.skip("TODO: enable for modin")
     left = integer_dataframe_1(library)
     right = integer_dataframe_2(library).rename({"b": "c"})
     result = left.join(right, left_on="a", right_on="a", how="left")
@@ -73,8 +71,6 @@ def test_join_outer(library: BaseHandler) -> None:  # pragma: no cover
 
 
 def test_join_two_keys(library: BaseHandler) -> None:
-    if library.name == "modin":
-        pytest.skip("TODO: enable for modin")
     left = integer_dataframe_1(library)
     right = integer_dataframe_2(library).rename({"b": "c"})
     result = left.join(right, left_on=["a", "b"], right_on=["a", "c"], how="left")

tests/integration/persistedness_test.py

@@ -55,8 +55,6 @@ def test_within_df_within_col_propagation(library: BaseHandler) -> None:
 
 
 def test_cross_df_propagation(library: BaseHandler) -> None:
-    if library.name == "modin":
-        pytest.skip("TODO: enable for modin")
     df1 = integer_dataframe_1(library)
     df2 = integer_dataframe_2(library)
     ns = df1.__dataframe_namespace__()