[DO NOT REVIEW] try to increase Java -Xmx memory for non-UTC time zone.

integration_tests/run_pyspark_from_build.sh

@@ -349,6 +349,8 @@ EOF
         echo "SUCCESS spark-shell smoke test"
     elif ((${#TEST_PARALLEL_OPTS[@]} > 0));
     then
+        # will set -Xmx=2g
+        export PYSP_TEST_spark_driver_memory=2g
         exec python "${RUN_TESTS_COMMAND[@]}" "${TEST_PARALLEL_OPTS[@]}" "${TEST_COMMON_OPTS[@]}"
     else
         # We set the GPU memory size to be a constant value even if only running with a parallelism of 1

integration_tests/src/main/python/aqe_test.py

@@ -16,6 +16,7 @@
 from pyspark.sql.functions import when, col, current_date, current_timestamp
 from pyspark.sql.types import *
 from asserts import assert_gpu_and_cpu_are_equal_collect, assert_cpu_and_gpu_are_equal_collect_with_capture
+from conftest import is_not_utc
 from data_gen import *
 from marks import ignore_order, allow_non_gpu
 from spark_session import with_cpu_session, is_databricks113_or_later
@@ -195,6 +196,7 @@ def do_it(spark):
 @ignore_order(local=True)
 @allow_non_gpu('BroadcastNestedLoopJoinExec', 'Cast', 'DateSub', *db_113_cpu_bnlj_join_allow)
 @pytest.mark.parametrize('join', joins, ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_aqe_join_reused_exchange_inequality_condition(spark_tmp_path, join):
     data_path = spark_tmp_path + '/PARQUET_DATA'
     def prep(spark):

integration_tests/src/main/python/arithmetic_ops_test.py

@@ -16,6 +16,7 @@
 import pytest
 
 from asserts import assert_gpu_and_cpu_are_equal_collect, assert_gpu_and_cpu_error, assert_gpu_fallback_collect, assert_gpu_and_cpu_are_equal_sql
+from conftest import is_not_utc
 from data_gen import *
 from marks import ignore_order, incompat, approximate_float, allow_non_gpu, datagen_overrides
 from pyspark.sql.types import *
@@ -985,6 +986,7 @@ def test_columnar_pow(data_gen):
             lambda spark : binary_op_df(spark, data_gen).selectExpr('pow(a, b)'))
 
 @pytest.mark.parametrize('data_gen', all_basic_gens + _arith_decimal_gens, ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_least(data_gen):
     num_cols = 20
     s1 = with_cpu_session(
@@ -1001,6 +1003,7 @@ def test_least(data_gen):
                 f.least(*command_args)))
 
 @pytest.mark.parametrize('data_gen', all_basic_gens + _arith_decimal_gens, ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_greatest(data_gen):
     num_cols = 20
     s1 = with_cpu_session(

integration_tests/src/main/python/array_test.py

@@ -16,7 +16,7 @@
 
 from asserts import assert_gpu_and_cpu_are_equal_collect, assert_gpu_and_cpu_are_equal_sql, assert_gpu_and_cpu_error, assert_gpu_fallback_collect
 from data_gen import *
-from conftest import is_databricks_runtime
+from conftest import is_databricks_runtime, is_not_utc
 from marks import incompat
 from spark_session import is_before_spark_313, is_before_spark_330, is_databricks113_or_later, is_spark_330_or_later, is_databricks104_or_later, is_spark_33X, is_spark_340_or_later, is_spark_330, is_spark_330cdh
 from pyspark.sql.types import *
@@ -103,11 +103,13 @@
 
 @pytest.mark.parametrize('data_gen', array_item_test_gens, ids=idfn)
 @pytest.mark.parametrize('index_gen', array_index_gens, ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_array_item(data_gen, index_gen):
     assert_gpu_and_cpu_are_equal_collect(
         lambda spark: two_col_df(spark, data_gen, index_gen).selectExpr('a[b]'))
 
 @pytest.mark.parametrize('data_gen', array_item_test_gens, ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_array_item_lit_ordinal(data_gen):
     assert_gpu_and_cpu_are_equal_collect(
         lambda spark: unary_op_df(spark, data_gen).selectExpr(
@@ -145,6 +147,7 @@ def test_array_item_with_strict_index(strict_index_enabled, index):
 
 # No need to test this for multiple data types for array. Only one is enough, but with two kinds of invalid index.
 @pytest.mark.parametrize('index', [-2, 100, array_neg_index_gen, array_out_index_gen], ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_array_item_ansi_fail_invalid_index(index):
     message = "SparkArrayIndexOutOfBoundsException" if (is_databricks104_or_later() or is_spark_330_or_later()) else "java.lang.ArrayIndexOutOfBoundsException"
     if isinstance(index, int):
@@ -171,6 +174,7 @@ def test_array_item_ansi_not_fail_all_null_data():
                          decimal_gen_32bit, decimal_gen_64bit, decimal_gen_128bit, binary_gen,
                          StructGen([['child0', StructGen([['child01', IntegerGen()]])], ['child1', string_gen], ['child2', float_gen]], nullable=False),
                          StructGen([['child0', byte_gen], ['child1', string_gen], ['child2', float_gen]], nullable=False)], ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_make_array(data_gen):
     (s1, s2) = with_cpu_session(
         lambda spark: gen_scalars_for_sql(data_gen, 2, force_no_nulls=not isinstance(data_gen, NullGen)))
@@ -183,6 +187,7 @@ def test_make_array(data_gen):
 
 
 @pytest.mark.parametrize('data_gen', single_level_array_gens, ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_orderby_array_unique(data_gen):
     assert_gpu_and_cpu_are_equal_sql(
         lambda spark : append_unique_int_col_to_df(spark, unary_op_df(spark, data_gen)),
@@ -212,6 +217,7 @@ def test_orderby_array_of_structs(data_gen):
 @pytest.mark.parametrize('data_gen', [byte_gen, short_gen, int_gen, long_gen,
                                       float_gen, double_gen,
                                       string_gen, boolean_gen, date_gen, timestamp_gen], ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_array_contains(data_gen):
     arr_gen = ArrayGen(data_gen)
     literal = with_cpu_session(lambda spark: gen_scalar(data_gen, force_no_nulls=True))
@@ -239,6 +245,7 @@ def test_array_contains_for_nans(data_gen):
 
 
 @pytest.mark.parametrize('data_gen', array_item_test_gens, ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_array_element_at(data_gen):
     assert_gpu_and_cpu_are_equal_collect(
         lambda spark: two_col_df(spark, data_gen, array_no_zero_index_gen).selectExpr(
@@ -303,6 +310,7 @@ def test_array_element_at_zero_index_fail(index, ansi_enabled):
 
 
 @pytest.mark.parametrize('data_gen', array_gens_sample, ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_array_transform(data_gen):
     def do_it(spark):
         columns = ['a', 'b',
@@ -337,6 +345,7 @@ def do_it(spark):
         string_gen, boolean_gen, date_gen, timestamp_gen, null_gen] + decimal_gens
 
 @pytest.mark.parametrize('data_gen', array_min_max_gens, ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_array_min_max(data_gen):
     assert_gpu_and_cpu_are_equal_collect(
             lambda spark : unary_op_df(spark, ArrayGen(data_gen)).selectExpr(
@@ -361,6 +370,7 @@ def test_array_concat_decimal(data_gen):
             'concat(a, a)')))
 
 @pytest.mark.parametrize('data_gen', orderable_gens + nested_gens_sample, ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_array_repeat_with_count_column(data_gen):
     cnt_gen = IntegerGen(min_val=-5, max_val=5, special_cases=[])
     cnt_not_null_gen = IntegerGen(min_val=-5, max_val=5, special_cases=[], nullable=False)
@@ -374,6 +384,7 @@ def test_array_repeat_with_count_column(data_gen):
 
 
 @pytest.mark.parametrize('data_gen', orderable_gens + nested_gens_sample, ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_array_repeat_with_count_scalar(data_gen):
     assert_gpu_and_cpu_are_equal_collect(
         lambda spark: unary_op_df(spark, data_gen).selectExpr(
@@ -403,6 +414,7 @@ def test_sql_array_scalars(query):
 
 
 @pytest.mark.parametrize('data_gen', all_basic_gens + nested_gens_sample, ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_get_array_struct_fields(data_gen):
     array_struct_gen = ArrayGen(
         StructGen([['child0', data_gen], ['child1', int_gen]]),
@@ -441,6 +453,7 @@ def do_it(spark):
 
 
 @pytest.mark.parametrize('data_gen', array_zips_gen, ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_arrays_zip(data_gen):
     gen = StructGen(
         [('a', data_gen), ('b', data_gen), ('c', data_gen), ('d', data_gen)], nullable=False)
@@ -473,6 +486,7 @@ def q1(spark):
 
 @incompat
 @pytest.mark.parametrize('data_gen', no_neg_zero_all_basic_gens + decimal_gens, ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 @pytest.mark.skipif(is_before_spark_313() or is_spark_330() or is_spark_330cdh(), reason="NaN equality is only handled in Spark 3.1.3+ and SPARK-39976 issue with null and ArrayIntersect in Spark 3.3.0")
 def test_array_intersect(data_gen):
     gen = StructGen(
@@ -514,6 +528,7 @@ def test_array_intersect_spark330(data_gen):
 @incompat
 @pytest.mark.parametrize('data_gen', no_neg_zero_all_basic_gens_no_nans + decimal_gens, ids=idfn)
 @pytest.mark.skipif(not is_before_spark_313(), reason="NaN equality is only handled in Spark 3.1.3+")
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_array_intersect_before_spark313(data_gen):
     gen = StructGen(
         [('a', ArrayGen(data_gen, nullable=True)),
@@ -534,6 +549,7 @@ def test_array_intersect_before_spark313(data_gen):
 @incompat
 @pytest.mark.parametrize('data_gen', no_neg_zero_all_basic_gens + decimal_gens, ids=idfn)
 @pytest.mark.skipif(is_before_spark_313(), reason="NaN equality is only handled in Spark 3.1.3+")
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_array_union(data_gen):
     gen = StructGen(
         [('a', ArrayGen(data_gen, nullable=True)),
@@ -554,6 +570,7 @@ def test_array_union(data_gen):
 @incompat
 @pytest.mark.parametrize('data_gen', no_neg_zero_all_basic_gens_no_nans + decimal_gens, ids=idfn)
 @pytest.mark.skipif(not is_before_spark_313(), reason="NaN equality is only handled in Spark 3.1.3+")
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_array_union_before_spark313(data_gen):
     gen = StructGen(
         [('a', ArrayGen(data_gen, nullable=True)),
@@ -574,6 +591,7 @@ def test_array_union_before_spark313(data_gen):
 @incompat
 @pytest.mark.parametrize('data_gen', no_neg_zero_all_basic_gens + decimal_gens, ids=idfn)
 @pytest.mark.skipif(is_before_spark_313(), reason="NaN equality is only handled in Spark 3.1.3+")
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_array_except(data_gen):
     gen = StructGen(
         [('a', ArrayGen(data_gen, nullable=True)),
@@ -594,6 +612,7 @@ def test_array_except(data_gen):
 @incompat
 @pytest.mark.parametrize('data_gen', no_neg_zero_all_basic_gens_no_nans + decimal_gens, ids=idfn)
 @pytest.mark.skipif(not is_before_spark_313(), reason="NaN equality is only handled in Spark 3.1.3+")
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_array_except_before_spark313(data_gen):
     gen = StructGen(
         [('a', ArrayGen(data_gen, nullable=True)),
@@ -614,6 +633,7 @@ def test_array_except_before_spark313(data_gen):
 @incompat
 @pytest.mark.parametrize('data_gen', no_neg_zero_all_basic_gens + decimal_gens, ids=idfn)
 @pytest.mark.skipif(is_before_spark_313(), reason="NaN equality is only handled in Spark 3.1.3+")
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_arrays_overlap(data_gen):
     gen = StructGen(
         [('a', ArrayGen(data_gen, nullable=True)),
@@ -635,6 +655,7 @@ def test_arrays_overlap(data_gen):
 @incompat
 @pytest.mark.parametrize('data_gen', no_neg_zero_all_basic_gens_no_nans + decimal_gens, ids=idfn)
 @pytest.mark.skipif(not is_before_spark_313(), reason="NaN equality is only handled in Spark 3.1.3+")
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_arrays_overlap_before_spark313(data_gen):
     gen = StructGen(
         [('a', ArrayGen(data_gen, nullable=True)),
@@ -672,6 +693,7 @@ def test_array_remove_scalar(data_gen):
                                       FloatGen(special_cases=_non_neg_zero_float_special_cases + [-0.0]), 
                                       DoubleGen(special_cases=_non_neg_zero_double_special_cases + [-0.0]),
                                       StringGen(pattern='[0-9]{1,5}'), boolean_gen, date_gen, timestamp_gen] + decimal_gens, ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_array_remove(data_gen):
     gen = StructGen(
         [('a', ArrayGen(data_gen, nullable=True)),
@@ -686,6 +708,7 @@ def test_array_remove(data_gen):
 
 
 @pytest.mark.parametrize('data_gen', [ArrayGen(sub_gen) for sub_gen in array_gens_sample], ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_flatten_array(data_gen):
     assert_gpu_and_cpu_are_equal_collect(
         lambda spark: unary_op_df(spark, data_gen).selectExpr('flatten(a)')

integration_tests/src/main/python/ast_test.py

@@ -15,6 +15,7 @@
 import pytest
 
 from asserts import assert_cpu_and_gpu_are_equal_collect_with_capture
+from conftest import is_not_utc
 from data_gen import *
 from marks import approximate_float, datagen_overrides
 from spark_session import with_cpu_session, is_before_spark_330
@@ -70,6 +71,7 @@ def assert_binary_ast(data_descr, func, conf={}):
     assert_gpu_ast(is_supported, lambda spark: func(binary_op_df(spark, data_gen)), conf=conf)
 
 @pytest.mark.parametrize('data_gen', [boolean_gen, byte_gen, short_gen, int_gen, long_gen, float_gen, double_gen, timestamp_gen, date_gen], ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_literal(spark_tmp_path, data_gen):
     # Write data to Parquet so Spark generates a plan using just the count of the data.
     data_path = spark_tmp_path + '/AST_TEST_DATA'
@@ -79,6 +81,7 @@ def test_literal(spark_tmp_path, data_gen):
                    func=lambda spark: spark.read.parquet(data_path).select(scalar))
 
 @pytest.mark.parametrize('data_gen', [boolean_gen, byte_gen, short_gen, int_gen, long_gen, float_gen, double_gen, timestamp_gen, date_gen], ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_null_literal(spark_tmp_path, data_gen):
     # Write data to Parquet so Spark generates a plan using just the count of the data.
     data_path = spark_tmp_path + '/AST_TEST_DATA'
@@ -232,6 +235,7 @@ def test_expm1(data_descr):
     assert_unary_ast(data_descr, lambda df: df.selectExpr('expm1(a)'))
 
 @pytest.mark.parametrize('data_descr', ast_comparable_descrs, ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_eq(data_descr):
     (s1, s2) = with_cpu_session(lambda spark: gen_scalars(data_descr[0], 2))
     assert_binary_ast(data_descr,
@@ -241,6 +245,7 @@ def test_eq(data_descr):
             f.col('a') == f.col('b')))
 
 @pytest.mark.parametrize('data_descr', ast_comparable_descrs, ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_ne(data_descr):
     (s1, s2) = with_cpu_session(lambda spark: gen_scalars(data_descr[0], 2))
     assert_binary_ast(data_descr,
@@ -250,6 +255,7 @@ def test_ne(data_descr):
             f.col('a') != f.col('b')))
 
 @pytest.mark.parametrize('data_descr', ast_comparable_descrs, ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_lt(data_descr):
     (s1, s2) = with_cpu_session(lambda spark: gen_scalars(data_descr[0], 2))
     assert_binary_ast(data_descr,
@@ -259,6 +265,7 @@ def test_lt(data_descr):
             f.col('a') < f.col('b')))
 
 @pytest.mark.parametrize('data_descr', ast_comparable_descrs, ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_lte(data_descr):
     (s1, s2) = with_cpu_session(lambda spark: gen_scalars(data_descr[0], 2))
     assert_binary_ast(data_descr,
@@ -268,6 +275,7 @@ def test_lte(data_descr):
             f.col('a') <= f.col('b')))
 
 @pytest.mark.parametrize('data_descr', ast_comparable_descrs, ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_gt(data_descr):
     (s1, s2) = with_cpu_session(lambda spark: gen_scalars(data_descr[0], 2))
     assert_binary_ast(data_descr,
@@ -277,6 +285,7 @@ def test_gt(data_descr):
             f.col('a') > f.col('b')))
 
 @pytest.mark.parametrize('data_descr', ast_comparable_descrs, ids=idfn)
+@pytest.mark.xfail(condition = is_not_utc(), reason = 'xfail non-UTC time zone tests because of https://github.com/NVIDIA/spark-rapids/issues/9653')
 def test_gte(data_descr):
     (s1, s2) = with_cpu_session(lambda spark: gen_scalars(data_descr[0], 2))
     assert_binary_ast(data_descr,