Type ignore SkipIf with Optional[str] message

This are, technically speaking, incorrect. However, messages are set
using the same condition as SkipIf, so there is nor practical issue
here.

python/pyspark/sql/tests/test_arrow.py

@@ -42,7 +42,7 @@
 
 @unittest.skipIf(
     not have_pandas or not have_pyarrow,
-    pandas_requirement_message or pyarrow_requirement_message)
+    pandas_requirement_message or pyarrow_requirement_message)  # type: ignore[arg-type]
 class ArrowTests(ReusedSQLTestCase):
 
     @classmethod
@@ -465,7 +465,7 @@ def test_createDataFrame_empty_partition(self):
 
 @unittest.skipIf(
     not have_pandas or not have_pyarrow,
-    pandas_requirement_message or pyarrow_requirement_message)
+    pandas_requirement_message or pyarrow_requirement_message)  # type: ignore[arg-type]
 class MaxResultArrowTests(unittest.TestCase):
     # These tests are separate as 'spark.driver.maxResultSize' configuration
     # is a static configuration to Spark context.

python/pyspark/sql/tests/test_dataframe.py

@@ -518,7 +518,7 @@ def _to_pandas(self):
         df = self.spark.createDataFrame(data, schema)
         return df.toPandas()
 
-    @unittest.skipIf(not have_pandas, pandas_requirement_message)
+    @unittest.skipIf(not have_pandas, pandas_requirement_message)  # type: ignore[arg-type]
     def test_to_pandas(self):
         import numpy as np
         pdf = self._to_pandas()
@@ -530,7 +530,7 @@ def test_to_pandas(self):
         self.assertEquals(types[4], np.object)  # datetime.date
         self.assertEquals(types[5], 'datetime64[ns]')
 
-    @unittest.skipIf(not have_pandas, pandas_requirement_message)
+    @unittest.skipIf(not have_pandas, pandas_requirement_message)  # type: ignore[arg-type]
     def test_to_pandas_with_duplicated_column_names(self):
         import numpy as np
 
@@ -543,7 +543,7 @@ def test_to_pandas_with_duplicated_column_names(self):
                 self.assertEquals(types.iloc[0], np.int32)
                 self.assertEquals(types.iloc[1], np.int32)
 
-    @unittest.skipIf(not have_pandas, pandas_requirement_message)
+    @unittest.skipIf(not have_pandas, pandas_requirement_message)  # type: ignore[arg-type]
     def test_to_pandas_on_cross_join(self):
         import numpy as np
 
@@ -569,7 +569,7 @@ def test_to_pandas_required_pandas_not_found(self):
             with self.assertRaisesRegexp(ImportError, 'Pandas >= .* must be installed'):
                 self._to_pandas()
 
-    @unittest.skipIf(not have_pandas, pandas_requirement_message)
+    @unittest.skipIf(not have_pandas, pandas_requirement_message)  # type: ignore[arg-type]
     def test_to_pandas_avoid_astype(self):
         import numpy as np
         schema = StructType().add("a", IntegerType()).add("b", StringType())\
@@ -581,7 +581,7 @@ def test_to_pandas_avoid_astype(self):
         self.assertEquals(types[1], np.object)
         self.assertEquals(types[2], np.float64)
 
-    @unittest.skipIf(not have_pandas, pandas_requirement_message)
+    @unittest.skipIf(not have_pandas, pandas_requirement_message)  # type: ignore[arg-type]
     def test_to_pandas_from_empty_dataframe(self):
         with self.sql_conf({"spark.sql.execution.arrow.pyspark.enabled": False}):
             # SPARK-29188 test that toPandas() on an empty dataframe has the correct dtypes
@@ -601,7 +601,7 @@ def test_to_pandas_from_empty_dataframe(self):
             dtypes_when_empty_df = self.spark.sql(sql).filter("False").toPandas().dtypes
             self.assertTrue(np.all(dtypes_when_empty_df == dtypes_when_nonempty_df))
 
-    @unittest.skipIf(not have_pandas, pandas_requirement_message)
+    @unittest.skipIf(not have_pandas, pandas_requirement_message)  # type: ignore[arg-type]
     def test_to_pandas_from_null_dataframe(self):
         with self.sql_conf({"spark.sql.execution.arrow.pyspark.enabled": False}):
             # SPARK-29188 test that toPandas() on a dataframe with only nulls has correct dtypes
@@ -629,7 +629,7 @@ def test_to_pandas_from_null_dataframe(self):
             self.assertEqual(types[7], np.object)
             self.assertTrue(np.can_cast(np.datetime64, types[8]))
 
-    @unittest.skipIf(not have_pandas, pandas_requirement_message)
+    @unittest.skipIf(not have_pandas, pandas_requirement_message)  # type: ignore[arg-type]
     def test_to_pandas_from_mixed_dataframe(self):
         with self.sql_conf({"spark.sql.execution.arrow.pyspark.enabled": False}):
             # SPARK-29188 test that toPandas() on a dataframe with some nulls has correct dtypes
@@ -657,7 +657,7 @@ def test_create_dataframe_from_array_of_long(self):
         df = self.spark.createDataFrame(data)
         self.assertEqual(df.first(), Row(longarray=[-9223372036854775808, 0, 9223372036854775807]))
 
-    @unittest.skipIf(not have_pandas, pandas_requirement_message)
+    @unittest.skipIf(not have_pandas, pandas_requirement_message)  # type: ignore[arg-type]
     def test_create_dataframe_from_pandas_with_timestamp(self):
         import pandas as pd
         from datetime import datetime
@@ -685,7 +685,7 @@ def test_create_dataframe_required_pandas_not_found(self):
                 self.spark.createDataFrame(pdf)
 
     # Regression test for SPARK-23360
-    @unittest.skipIf(not have_pandas, pandas_requirement_message)
+    @unittest.skipIf(not have_pandas, pandas_requirement_message)  # type: ignore[arg-type]
     def test_create_dataframe_from_pandas_with_dst(self):
         import pandas as pd
         from pandas.util.testing import assert_frame_equal
@@ -889,7 +889,7 @@ def test_query_execution_listener_on_collect(self):
 
     @unittest.skipIf(
         not have_pandas or not have_pyarrow,
-        pandas_requirement_message or pyarrow_requirement_message)
+        pandas_requirement_message or pyarrow_requirement_message)  # type: ignore[arg-type]
     def test_query_execution_listener_on_collect_with_arrow(self):
         with self.sql_conf({"spark.sql.execution.arrow.pyspark.enabled": True}):
             self.assertFalse(

python/pyspark/sql/tests/test_pandas_cogrouped_map.py

@@ -33,7 +33,7 @@
 
 @unittest.skipIf(
     not have_pandas or not have_pyarrow,
-    pandas_requirement_message or pyarrow_requirement_message)
+    pandas_requirement_message or pyarrow_requirement_message)  # type: ignore[arg-type]
 class CogroupedMapInPandasTests(ReusedSQLTestCase):
 
     @property

python/pyspark/sql/tests/test_pandas_grouped_map.py

@@ -41,7 +41,7 @@
 
 @unittest.skipIf(
     not have_pandas or not have_pyarrow,
-    pandas_requirement_message or pyarrow_requirement_message)
+    pandas_requirement_message or pyarrow_requirement_message)  # type: ignore[arg-type]
 class GroupedMapInPandasTests(ReusedSQLTestCase):
 
     @property

python/pyspark/sql/tests/test_pandas_map.py

@@ -27,7 +27,7 @@
 
 @unittest.skipIf(
     not have_pandas or not have_pyarrow,
-    pandas_requirement_message or pyarrow_requirement_message)
+    pandas_requirement_message or pyarrow_requirement_message)  # type: ignore[arg-type]
 class MapInPandasTests(ReusedSQLTestCase):
 
     @classmethod

python/pyspark/sql/tests/test_pandas_udf.py

@@ -28,7 +28,7 @@
 
 @unittest.skipIf(
     not have_pandas or not have_pyarrow,
-    pandas_requirement_message or pyarrow_requirement_message)
+    pandas_requirement_message or pyarrow_requirement_message)  # type: ignore[arg-type]
 class PandasUDFTests(ReusedSQLTestCase):
 
     def test_pandas_udf_basic(self):

python/pyspark/sql/tests/test_pandas_udf_grouped_agg.py

@@ -35,7 +35,7 @@
 
 @unittest.skipIf(
     not have_pandas or not have_pyarrow,
-    pandas_requirement_message or pyarrow_requirement_message)
+    pandas_requirement_message or pyarrow_requirement_message)  # type: ignore[arg-type]
 class GroupedAggPandasUDFTests(ReusedSQLTestCase):
 
     @property

python/pyspark/sql/tests/test_pandas_udf_scalar.py

@@ -46,7 +46,7 @@
 
 @unittest.skipIf(
     not have_pandas or not have_pyarrow,
-    pandas_requirement_message or pyarrow_requirement_message)
+    pandas_requirement_message or pyarrow_requirement_message)  # type: ignore[arg-type]
 class ScalarPandasUDFTests(ReusedSQLTestCase):
 
     @classmethod
@@ -1095,7 +1095,7 @@ def f3i(it):
             self.assertEquals(expected, df1.collect())
 
     # SPARK-24721
-    @unittest.skipIf(not test_compiled, test_not_compiled_message)
+    @unittest.skipIf(not test_compiled, test_not_compiled_message)  # type: ignore[arg-type]
     def test_datasource_with_udf(self):
         # Same as SQLTests.test_datasource_with_udf, but with Pandas UDF
         # This needs to a separate test because Arrow dependency is optional

python/pyspark/sql/tests/test_pandas_udf_typehints.py

@@ -34,7 +34,7 @@
 
 @unittest.skipIf(
     not have_pandas or not have_pyarrow,
-    pandas_requirement_message or pyarrow_requirement_message)
+    pandas_requirement_message or pyarrow_requirement_message)  # type: ignore[arg-type]
 class PandasUDFTypeHintsTests(ReusedSQLTestCase):
     def test_type_annotation_scalar(self):
         def func(col: pd.Series) -> pd.Series:

python/pyspark/sql/tests/test_pandas_udf_window.py

@@ -31,7 +31,7 @@
 
 @unittest.skipIf(
     not have_pandas or not have_pyarrow,
-    pandas_requirement_message or pyarrow_requirement_message)
+    pandas_requirement_message or pyarrow_requirement_message)  # type: ignore[arg-type]
 class WindowPandasUDFTests(ReusedSQLTestCase):
     @property
     def data(self):

python/pyspark/sql/tests/test_udf.py

@@ -356,7 +356,7 @@ def test_udf_registration_returns_udf(self):
             df.select(add_four("id").alias("plus_four")).collect()
         )
 
-    @unittest.skipIf(not test_compiled, test_not_compiled_message)
+    @unittest.skipIf(not test_compiled, test_not_compiled_message)  # type: ignore[arg-type]
     def test_register_java_function(self):
         self.spark.udf.registerJavaFunction(
             "javaStringLength", "test.org.apache.spark.sql.JavaStringLength", IntegerType())
@@ -373,7 +373,7 @@ def test_register_java_function(self):
         [value] = self.spark.sql("SELECT javaStringLength3('test')").first()
         self.assertEqual(value, 4)
 
-    @unittest.skipIf(not test_compiled, test_not_compiled_message)
+    @unittest.skipIf(not test_compiled, test_not_compiled_message)  # type: ignore[arg-type]
     def test_register_java_udaf(self):
         self.spark.udf.registerJavaUDAF("javaUDAF", "test.org.apache.spark.sql.MyDoubleAvg")
         df = self.spark.createDataFrame([(1, "a"), (2, "b"), (3, "a")], ["id", "name"])
@@ -560,7 +560,7 @@ def test_nonparam_udf_with_aggregate(self):
         self.assertEqual(rows, [Row(_1=1, _2=2, a=u'const_str')])
 
     # SPARK-24721
-    @unittest.skipIf(not test_compiled, test_not_compiled_message)
+    @unittest.skipIf(not test_compiled, test_not_compiled_message)  # type: ignore[arg-type]
     def test_datasource_with_udf(self):
         from pyspark.sql.functions import lit, col