What is the best way of reading/writing binary input/output files with MapReduce?

In all samples I've seen so far, mapreduce apps take text files as input and write text as output.

I'd like my app to 开发者_开发知识库read objects from the binary file and write objects back to output file.

What is the best way to do that in MapReduce?

I'm writing the app in java

---

SequenceFile provides a persistent data structure for binary key-value pairs. You can find more information in the below URL.

http://wiki.apache.org/hadoop/SequenceFile

http://hadoop.apache.org/common/docs/current/api/org/apache/hadoop/io/SequenceFile.html