使用MapReduce统计微博关注数据

MapReduce的思想可以参考http://blog.csdn.net/mrbcy/article/details/60139191

总体思路

输入	输出
Mapper	用户1 用户2
Reducer	<用户名,1/-1>

上表中的-1代表入度，1代表出度

实现代码

代码内容如下：

package tech.mrbcy.bigdata.weiborelation;

import java.io.IOException;

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.IntWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.Mapper;
import org.apache.hadoop.mapreduce.Reducer;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;

public class WeiboRelation {
public static class RelationMapper
extends Mapper<Object, Text, Text, IntWritable>{
private boolean flag = true;
public void map(Object key, Text value, Context context
) throws IOException, InterruptedException {
//10457   104594
String valStr = value.toString().replace("  ", ",");
String[] users = valStr.split(",");
if(flag){
System.out.println(users[0] + "," + users[1]);
flag = false;
}
IntWritable inDegree = new IntWritable(-1);
IntWritable outDegree = new IntWritable(1);
context.write(new Text(users[0]), outDegree);
context.write(new Text(users[1]), inDegree);

}

}

public static class RelationReducer
extends Reducer<Text,IntWritable,Text,Text> {

public void reduce(Text key, Iterable<IntWritable> values,
Context context
) throws IOException, InterruptedException {
int inDegree = 0;
int outDegree = 0;

for (IntWritable val : values) {
if(val.get() > 0){
// 出度
outDegree += val.get();
}else{
// 入度
inDegree += -val.get();
}
}
String outString = String.format("%d    %d", inDegree,outDegree);
context.write(key, new Text(outString));
}
}

public static void main(String[] args) throws Exception {
Configuration conf = new Configuration();
Job job = Job.getInstance(conf,"statistic weibo user relation");
job.setJarByClass(WeiboRelation.class);
job.setMapperClass(RelationMapper.class);
job.setReducerClass(RelationReducer.class);
job.setMapOutputKeyClass(Text.class);
job.setMapOutputValueClass(IntWritable.class);
job.setOutputKeyClass(Text.class);
job.setOutputValueClass(Text.class);

FileInputFormat.addInputPath(job, new Path(args[0]));
FileOutputFormat.setOutputPath(job, new Path(args[1]));
System.exit(job.waitForCompletion(true) ? 0 : 1);
}
}

运行任务

首先把工程打成一个jar包导出。这一步我是用MyEclipse完成的。

然后把relation.txt拷贝到/root/homework/week2下面。

然后使用下面的命令在HDFS中建立输入文件夹，并确保输出文件夹不存在：

hadoop fs -mkdir -p /wbrelation/input
hadoop fs -rm -r /wbrelation/output

将input.txt上传到HDFS中：

hadoop fs -put /root/homework/week2/relation.txt /wbrelation/input

使用下面的命令运行MapReduce程序。

hadoop jar /root/homework/week2/weiborelation.jar tech.mrbcy.bigdata.weiborelation.WeiboRelation /wbrelation/input /wbrelation/output

等待执行结束后，使用下面的命令查看结果：

hadoop fs -get /wbrelation/output/part-r-00000
mv part-r-00000 wbrelation.out
more wbrelation.out

输出结果如下：

1000003374      1    2
1000060787      0    4
1000079167      1    1
1000085444      2    0
1000092382      1    0
1000094052      0    1
1000095500      1    0
1000097914      2    8
1000098257      0    1
1000106075      3    0
1000117663      3    1
1000131684      0    1
1000136955      0    1
1000144332      0    2
1000148680      2    0
1000158915      0    1
1000166111      4    30
1000172520      0    1
1000195387      3    1
1000196715      0    1
1000213720      0    1
1000216284      0    1
1000240153      0    1
1000242321      0    1
1000256983      0    1
1000258991      1    0
1000269641      0    1
1000273720
b215
0    1
1000281111      1    1
1000284465      0    1
1000305443      6    2