ml.shifu.guagua.mapreduce
类 GuaguaWritableAdapter<W extends org.apache.hadoop.io.Writable>

java.lang.Object
  ml.shifu.guagua.io.HaltBytable
      ml.shifu.guagua.mapreduce.GuaguaWritableAdapter<W>

所有已实现的接口：

ml.shifu.guagua.io.Bytable

public class GuaguaWritableAdapter<W extends org.apache.hadoop.io.Writable>
extends ml.shifu.guagua.io.HaltBytable

GuaguaWritableAdapter is used to adapt hadoop io to Bytable interface.

The reason we use our own writable interface is that we'd like to not depend on hadoop io on guagua core service.

By using this adapter we don't need to implement such as IntWritable, LongWritable, Text and other common writable implementations.

Notice: If you use GuaguaWritableAdapter as your result class, you should provide the inner writable class name to GuaguaConstants.GUAGUA_MASTER_RESULT_CLASS or GuaguaConstants.GUAGUA_WORKER_RESULT_CLASS setting. That means to set '-mr' or 'wr' to hadoop writable class name. Here is a sample:

 hadoop jar ../lib/guagua-mapreduce-0.0.1-SNAPSHOT.jar \
     ml.shifu.guagua.mapreduce.GuaguaMapReduceClient   \
     -libjars ../lib/guagua-mapreduce-examples-0.0.1-SNAPSHOT.jar,../lib/guava-r09-jarjar.jar,../lib/encog-core-3.1.0.jar,../lib/guagua-mapreduce-0.0.1-SNAPSHOT.jar,../lib/guagua-core-0.0.1-SNAPSHOT.jar,../lib/zookeeper-3.4.5.jar \
     -i <input>  \
     -z <zkhost:zkport>  \
     -w ml.shifu.guagua.example.sum.SumWorker  \
     -m ml.shifu.guagua.example.sum.SumMaster  \
     -c 10 \
     -n "Guagua Sum Master-Workers Job" \
     -mr org.apache.hadoop.io.LongWritable \
     -wr org.apache.hadoop.io.LongWritable \
     $queue_opts
 

If you set your own main class, not use GuaguaMapReduceClient, you can set the result class like this.

 conf.set(GuaguaConstants.GUAGUA_MASTER_RESULT_CLASS, LongWritable.class.getName());
         conf.set(GuaguaConstants.GUAGUA_WORKER_RESULT_CLASS, LongWritable.class.getName());
 
 

另请参见：

GuaguaWritableSerializer

构造方法摘要

GuaguaWritableAdapter(W writable)
Contructor with Hadoop Writable setting.

方法摘要

void	doReadFields(DataInput in)
void	doWrite(DataOutput out)
boolean	equals(Object obj)
W	getWritable()
int	hashCode()
void	setWritable(W writable)
String	toString()

从类 ml.shifu.guagua.io.HaltBytable 继承的方法

isHalt, readFields, setHalt, write

从类 java.lang.Object 继承的方法

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

构造方法详细信息

GuaguaWritableAdapter

public GuaguaWritableAdapter(W writable)

Contructor with Hadoop Writable setting.

参数：

writable - Hadoop Writable instance

方法详细信息

doWrite

public void doWrite(DataOutput out)
             throws IOException

指定者：

类 ml.shifu.guagua.io.HaltBytable 中的 doWrite

抛出：

IOException

doReadFields

public void doReadFields(DataInput in)
                  throws IOException

指定者：

类 ml.shifu.guagua.io.HaltBytable 中的 doReadFields

抛出：

IOException

getWritable

public W getWritable()

setWritable

public void setWritable(W writable)

hashCode

public int hashCode()

覆盖：

类 Object 中的 hashCode

equals

public boolean equals(Object obj)

覆盖：

类 Object 中的 equals

toString

public String toString()

覆盖：

类 Object 中的 toString