com.bigdata.service.mapred.tasks
Class CountKeywords

java.lang.Object
  com.bigdata.service.mapred.AbstractReduceTask
      com.bigdata.service.mapred.tasks.CountKeywords

All Implemented Interfaces:

IReduceTask, ITask, Serializable

public class CountKeywords
extends AbstractReduceTask

Summarizes tuples of the form {key, term}.

Note that many terms may be conflated into the same Unicode sort key depending on the collator that you are using. This task just deserializes the 1st term entry for each distinct key. If you want some consistency in the reported terms, then you should normalize the terms in your map task.

Version:

$Id: CountKeywords.java 2265 2009-10-26 12:51:06Z thompsonbry $

Author:

Bryan Thompson

See Also:

ExtractKeywords, Serialized Form

Field Summary

Fields inherited from class com.bigdata.service.mapred.AbstractReduceTask

uuid

Constructor Summary

CountKeywords(UUID uuid, UUID dataService)

Method Summary

void	reduce(byte[] key, Iterator<byte[]> vals) Each reduce task will be presented with a series of key-value pairs in key order.

Methods inherited from class com.bigdata.service.mapred.AbstractReduceTask

getDataService, getUUID

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

CountKeywords

public CountKeywords(UUID uuid,
                     UUID dataService)

Method Detail

reduce

public void reduce(byte[] key,
                   Iterator<byte[]> vals)
            throws Exception

Description copied from interface: IReduceTask

Each reduce task will be presented with a series of key-value pairs in key order. However, the keys will be distributed across the N reduce tasks by the used defined hash function, so this is NOT a total ordering over the intermediate keys.

Note: This method is never invoked for a key for which there are no values.

Parameters:

key - A key.

vals - An iterator that will visit the set of values for that key.

Throws:

Exception