Map-only Jobs

Pangool comes with a convenient builder for Map-only jobs that accepts instances of Mappers instead of static classes. It is called MapOnlyJobBuilder. You can see the “Grep” example in the examples/ package. In this example we are creating a convenient map-only job with a Mapper that accepts a regex and we are doing it in only 3 lines:

 MapOnlyJobBuilder b = new MapOnlyJobBuilder(conf);
 b.setOutput(new Path(output), new HadoopOutputFormat(TextOutputFormat.class), Text.class, NullWritable.class);
 b.addInput(new Path(input), new HadoopInputFormat(TextInputFormat.class), new GrepHandler(regex));

Map-only Jobs have the same available features than Pangool MapReduce Jobs: named outputs, binary tuple serialization, etc.

Next: Rollup »