hadoop 输出MultipleOutputs学习及应用情境

MultipleOutputs可以轻易的将输出数据输出为多个。

案例一：writing to additional outputs other than the job default output.

案例二：to write data to different files provided by user

举例：

 * Usage pattern for job submission:
 * <pre>
 *
 * Job job = new Job();
 *
 * FileInputFormat.setInputPath(job, inDir);
 * FileOutputFormat.setOutputPath(job, outDir);
 *
 * job.setMapperClass(MOMap.class);
 * job.setReducerClass(MOReduce.class);
 * ...
 *
 * // Defines additional single text based output 'text' for the job
 * MultipleOutputs.addNamedOutput(job, "text", TextOutputFormat.class, LongWritable.class, Text.class);
 *
 * // Defines additional sequence-file based output 'sequence' for the job
 * MultipleOutputs.addNamedOutput(job, "seq", SequenceFileOutputFormat.class, LongWritable.class, Text.class);
 * ...
 *
 * job.waitForCompletion(true);
 * ...

 * </pre>
 * <p>
 * Usage in Reducer:
 * <pre>
 * <K, V> String generateFileName(K k, V v) {
 *   return k.toString() + "_" + v.toString();
 * }
 *
 * public class MOReduce extends
 *   Reducer&lt;WritableComparable, Writable,WritableComparable, Writable&gt; {
 * private MultipleOutputs mos;
 * public void setup(Context context) {
 * ...
 * mos = new MultipleOutputs(context);
 * }
 *
 * public void reduce(WritableComparable key, Iterator&lt;Writable&gt; values,
 * Context context)
 * throws IOException {
 * ...
 * mos.write("text", , key, new Text("Hello"));
 * mos.write("seq", LongWritable(1), new Text("Bye"), "seq_a");
 * mos.write("seq", LongWritable(2), key, new Text("Chau"), "seq_b");
 * mos.write(key, new Text("value"), generateFileName(key, new Text("value")));
 * ...
 * }
 *
 * public void cleanup(Context) throws IOException {
 * mos.close();
 * ...
 * }
 *
 * }
 * </pre>