Because breaks define divisions rather than breaks, all indexable keys will have a bucket.
Because breaks define divisions rather than breaks, all indexable keys will have a bucket. Keys that are far out of bounds will be assigned to either first or last partition.
Uses KeyIndex to partition an RDD in memory, giving its records some spatial locality. When persisting an RDD partitioned by this partitioner we can safely assume that all records contributing to the same SFC index will reside in one partition.