How do you check the (Key, Value) of a MR Hive Job?


I wonder how hive implement my SQL to MR job. To be specific, it there a way to check all the exact input/output K,V records of a MapTask/Reduce on Hadoop Web log ?

If I write MR application by my own, I can manually print each KV to log, so that I can check them on the stdout file. But how can I do this with Hive MR app ?