Top N values by Hadoop Map Reduce code

I am very new in hadoop world and struggling to achieve one simple task.

Can anybody please tell me how to get top n values for word count example by using only Map reduce code technique?

I do not want to use any hadoop command for this simple task.