Passed

org.apache.spark.sql.catalyst.expressions.aggregate.ApproximatePercentileSuite.class PercentileDigest, makes sure the memory foot print is bounded