[BUG]Dataproc test_hash_groupby_approx_percentile_long_repeated_keys failures

[tgravescs]

Describe the bug
11:12:07 =================================== FAILURES ===================================
11:12:07 ________ test_hash_groupby_approx_percentile_long_repeated_keys[false] _________
11:12:07
11:12:07 aqe_enabled = 'false'
11:12:07
11:12:07 @pytest.mark.parametrize('aqe_enabled', ['false', 'true'], ids=idfn)
11:12:07 def test_hash_groupby_approx_percentile_long_repeated_keys(aqe_enabled):
11:12:07 conf = copy_and_update(_approx_percentile_conf, {'spark.sql.adaptive.enabled': aqe_enabled})
11:12:07 > compare_percentile_approx(
11:12:07 lambda spark: gen_df(spark, [('k', RepeatSeqGen(LongGen(), length=20)),
11:12:07 ('v', LongRangeGen())], length=100),
11:12:07 [0.05, 0.25, 0.5, 0.75, 0.95], conf)
11:12:07
11:12:07 integration_tests/src/main/python/hash_aggregate_test.py:1160:
...
E assert None is not None
11:12:07
11:12:07 integration_tests/src/main/python/hash_aggregate_test.py:1263: AssertionError

[tgravescs]

duplicate of #4060