Ensure that partitions are read back in in the same order they were written; should be unnecessary as of Hadoop 2.8 / 3.x.
Ensure that partitions are read back in in the same order they were written; should be unnecessary as of Hadoop 2.8 / 3.x. See https://issues.apache.org/jira/browse/HADOOP-10798
SequenceFileInputFormat that guarantees loading the same splits it was written with.