An implementation of Hadoop's mapred and mapreduce input and output formats for ORC files. They use the core reader and writer, but present the data to the user in Writable objects.