Okay, I found a solution to my second problem. Beside of using collector.collect(keyText, value) to generate the multiple output, I still have to commit the same output to the output output.collect(key,value), only in this way the data can be exchanged between the Combiner and the Reducer.
But the first question sill remains, how to set up combiner when input is <K2,V2> while output is <K3,V3>? -- View this message in context: http://old.nabble.com/Combiner-and-MultipleOutputs-in-Mapreduce-tp29893459p29893493.html Sent from the Hadoop core-dev mailing list archive at Nabble.com.
