Changing The Output File Prefix Of Hadoop MapReduce Job

Your Hadoop job can have multiple reducers and each reducer will create a file by default with prefix part-r-xxxxx. The first reducer will create a file as part-r-00000 and second reducer will create a file as part-r-00001

What if you don’t like the default prefix “part” and would like to change the prefix?

Changing The Output File Prefix

We are in luck, Hadoop has mapreduce.output.basename property which we can use to set out the prefix of the output. You can use the property like below while executing the MapReduce job.

hadoop jar ~/MaxClosePriceTool-1.0-ARG.jar com.hirw.maxcloseprice.MaxClosePriceTool -D mapreduce.output.basename=custombase /user/hirw/input/stocks output/mapreduce/stocks

For the above command to work, your driver program needs to implement the Tool interface. Here is a sample.

public class MaxClosePriceTool extends Configured implements Tool {

@Override
public int run(String[] args) throws Exception {

.....

}

 }

Here is the output –