Hadoop C++ Pipes中context常见成员函数的作用

getJobConf¶

Get the JobConf for the current task

getInputKey¶

Get the current key

getInputValue¶

Get the current value

In the reducer, context.getInputValue is not available till context.nextValue is called !

progress¶

This method simply phones home to the NameNode, letting it know that the mapper or reducer is still working and has not died or zombified.

setStatus¶

The status message can be found in the hadoop*tasktracker*.log and in the web interface as "Status".

1 context.setStatus("Teke-lili");

getCounter¶

The counter will be displayed in the Web interface. You will have to get it once on init of the class.

nextValue¶

Iterate over the values. Important: The key will be the same all the time !

context.getInputValue is not available till context.nextValue is called 

例子：

假设输入文件是hello.txt

内容为:

hello world 

hello bupt

程序为：

#include "hadoop/Pipes.hh" 
#include "hadoop/TemplateFactory.hh" 
#include "hadoop/StringUtils.hh" 

const std::string WORDCOUNT = "WORDCOUNT"; 
const std::string INPUT_WORDS = "INPUT_WORDS"; 
const std::string OUTPUT_WORDS = "OUTPUT_WORDS"; 

class WordCountMap: public HadoopPipes::Mapper { 
// Mapper类 
public: 
	HadoopPipes::TaskContext::Counter* inputWords; 
	WordCountMap(HadoopPipes::TaskContext& context) 
	{ 
		inputWords = context.getCounter(WORDCOUNT, INPUT_WORDS); 
	} 
	void map(HadoopPipes::MapContext& context) 
	{ 
		std::vector<std::string> words = HadoopUtils::splitString(context.getInputValue(), " "); // 按空格进行单词分割 
		for(unsigned int i=0; i < words.size(); ++i) 
		{ 
			context.emit(words[i], "1"); // 单词作为key,value为1 
		} 
		context.incrementCounter(inputWords, words.size()); // 向map-reduce提交进度信息 
	} 
}; 

class WordCountReduce: public HadoopPipes::Reducer 
{ // reduce类 
public: 
	HadoopPipes::TaskContext::Counter* outputWords; 
	WordCountReduce(HadoopPipes::TaskContext& context) 
	{ 
		outputWords = context.getCounter(WORDCOUNT, OUTPUT_WORDS); 
	} 
	void reduce(HadoopPipes::ReduceContext& context) 
	{ 
		int sum = 0; 
		while (context.nextValue()) 
		{ 
			sum += HadoopUtils::toInt(context.getInputValue()); // 统计单词出现的次数 
		} 
		context.emit(context.getInputKey(), HadoopUtils::toString(sum)); // 输出结果 
		context.incrementCounter(outputWords, 1); 
	} 
}; 

int main(int argc, char *argv[]) 
{ 
	return HadoopPipes::runTask(HadoopPipes::TemplateFactory<WordCountMap, WordCountReduce>()); // 运行任务 
}

一。MapContext：

内容为：

key->value

(1,hello word)  注：这里的1是该行的偏移量，具体值不一定是这个

(2,hello bupt)

getInpuptValue() 可以得到一行的value，例如头一次调用将得到：hello world

emit()

将以下内容写入

(hello,1)

(world,1)

(hell0,1)

(bupt,1)

二。ReduceContext：

以上一步的内容为输入，经过MapReduce框架处理以后得到，内容为：

(hello,[1,1])  注：这里已经将key相同的value放到了一块

(world,1)

(bupt,1)

context.nextValue() 将会前进到特定key的下一个Value