in spark-job/src/main/java/org/apache/cassandra/diff/Differ.java [241:252]
static Predicate<PartitionKey> shouldIncludePartition(final UUID jobId, final double partitionSamplingProbability) {
if (partitionSamplingProbability > 1 || partitionSamplingProbability <= 0) {
logger.error("Invalid partition sampling property {}, it should be between 0 and 1", partitionSamplingProbability);
throw new IllegalArgumentException("Invalid partition sampling property, it should be between 0 and 1");
}
if (partitionSamplingProbability == 1) {
return partitionKey -> true;
} else {
final Random random = new Random(jobId.hashCode());
return partitionKey -> random.nextDouble() <= partitionSamplingProbability;
}
}