On Jun 15, 2009, at 9:44 AM, Kunsheng Chen wrote:
I got some question on my situation. My first Map/Reduce pass would finish all work I need. The output results would be divided into 2 groups, and each result has a priority in its group.
You should look at the MultipleOutputs in org.apache.hadoop.mapred.lib. It lets you write multiple streams out of your reducer.
-- Owen
