Disable testTableSampleBernoulli in BigQuery

[ebyhr]

Description

The test is slow and increases the possibility of CI failures due to BigQuery SLA because it contains 100 executions.
Disabling this test should be safe since the connector doesn't support SAMPLE pushdown.
I think we can make this test conditional with TestingConnectorBehavior in a follow-up.

trino/testing/trino-testing/src/main/java/io/trino/testing/AbstractTestQueries.java

Lines 456 to 476 in 90c48bb

	@Test
	public void testTableSampleBernoulli()
	{
	DescriptiveStatistics stats = new DescriptiveStatistics();
	int total = computeExpected("SELECT orderkey FROM orders", ImmutableList.of(BIGINT)).getMaterializedRows().size();
	for (int i = 0; i < 100; i++) {
	List<MaterializedRow> values = computeActual("SELECT orderkey FROM orders TABLESAMPLE BERNOULLI (50)").getMaterializedRows();
	assertThat(values.size())
	.describedAs("TABLESAMPLE produced duplicate rows")
	.isEqualTo(ImmutableSet.copyOf(values).size());
	stats.addValue(values.size() * 1.0 / total);
	}
	double mean = stats.getGeometricMean();
	assertThat(mean > 0.45 && mean < 0.55)
	.describedAs(format("Expected mean sampling rate to be ~0.5, but was %s", mean))
	.isTrue();
	}

Release notes

(x) This is not user-visible or is docs only, and no release notes are required.