Hadoop 自定义序列化编程

一 自定义序列化需求

二 MapReduce代码编写

1 自定义序列化类

package com.cakin.hadoop.mr;
import java.io.DataInput;
import java.io.DataOutput;
import java.io.IOException;
import org.apache.hadoop.io.WritableComparable;
public class UserWritable implements WritableComparable<UserWritable> {
     private Integer id;
     private Integer income;
     private Integer expenses;
     private Integer sum;
     
     public void write(DataOutput out) throws IOException {
           // TODO Auto-generated method stub
           out.writeInt(id);
           out.writeInt(income);
           out.writeInt(expenses);
           out.writeInt(sum);
     }
     public void readFields(DataInput in) throws IOException {
           // TODO Auto-generated method stub
           this.id=in.readInt();
           this.income=in.readInt();
           this.expenses=in.readInt();
           this.sum=in.readInt();          
     }
     
     public Integer getId() {
           return id;
     }
     public UserWritable setId(Integer id) {
           this.id = id;
           return this;
     }
     public Integer getIncome() {
           return income;
     }
     public UserWritable setIncome(Integer income) {
           this.income = income;
           return this;
     }
     public Integer getExpenses() {
           return expenses;
     }
     public UserWritable setExpenses(Integer expenses) {
           this.expenses = expenses;
           return this;
     }
     public Integer getSum() {
           return sum;
     }
     public UserWritable setSum(Integer sum) {
           this.sum = sum;
           return this;
     }
     public int compareTo(UserWritable o) {
           // TODO Auto-generated method stub
           return this.id>o.getId()?1:-1;
     }
     @Override
     public String toString() {
           return id + "\t"+income+"\t"+expenses+"\t"+sum;
     }
     
}

2 编写MapReduce

package com.cakin.hadoop.mr;
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.LongWritable;
import org.apache.hadoop.io.NullWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.Mapper;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
import org.apache.hadoop.mapreduce.Reducer;
/*
* 测试数据
* 用户id        收入        支出
* 1        1000    0
* 2        500        300
* 1        2000    1000
* 2        500        200
*
* 需求：
* 用户id        总收入    总支出    总的余额
* 1        3000    1000    2000
* 2        1000    500        500
* */
public class CountMapReduce {
    public static class CountMapper extends Mapper<LongWritable,Text,IntWritable,UserWritable>
    {
         private UserWritable userWritable =new UserWritable();
         private IntWritable id =new IntWritable();
         @Override
         protected void map(LongWritable key,Text value,
                 Mapper<LongWritable,Text,IntWritable,UserWritable>.Context context) throws IOException, InterruptedException{
             String line = value.toString();
             String[] words = line.split("\t");
             if(words.length ==3)
             {
                 userWritable.setId(Integer.parseInt(words[0]))
                 .setIncome(Integer.parseInt(words[1]))
                 .setExpenses(Integer.parseInt(words[2]))
                 .setSum(Integer.parseInt(words[1])-Integer.parseInt(words[2]));
                 id.set(Integer.parseInt(words[0]));
             }
             context.write(id, userWritable);
         }
    }
    public static class CountReducer extends Reducer<IntWritable,UserWritable,UserWritable,NullWritable>
    {
        /*
         * 输入数据
         * <1,{[1,1000,0,1000],[1,2000,1000,1000]}>
         * <2,[2,500,300,200],[2,500,200,300]>
         *
         * */
        
         private UserWritable userWritable = new UserWritable();
         private NullWritable n = NullWritable.get();
         protected void reduce(IntWritable key,Iterable<UserWritable> values,
                 Reducer<IntWritable,UserWritable,UserWritable,NullWritable>.Context context) throws IOException, InterruptedException{
             Integer income=0;
             Integer expenses = 0;
             Integer sum =0;
             for(UserWritable u:values)
             {
                 income += u.getIncome();
                 expenses+=u.getExpenses();
             }
             sum = income - expenses;
             userWritable.setId(key.get())
             .setIncome(income)
             .setExpenses(expenses)
             .setSum(sum);
             context.write(userWritable, n);
         }
    }
    public static void main(String[] args) throws IllegalArgumentException, IOException, ClassNotFoundException, InterruptedException {
        Configuration conf=new Configuration();
        /*
         * 集群中节点都有配置文件
        conf.set("mapreduce.framework.name.", "yarn");
        conf.set("yarn.resourcemanager.hostname", "mini1");
        */
        Job job=Job.getInstance(conf,"countMR");
        //jar包在哪里,现在在客户端，传递参数
        //任意运行，类加载器知道这个类的路径，就可以知道jar包所在的本地路径
        job.setJarByClass(CountMapReduce.class);
        //指定本业务job要使用的mapper/Reducer业务类
        job.setMapperClass(CountMapper.class);
        job.setReducerClass(CountReducer.class);
        //指定mapper输出数据的kv类型
        job.setMapOutputKeyClass(IntWritable.class);
        job.setMapOutputValueClass(UserWritable.class);
        //指定最终输出的数据kv类型
        job.setOutputKeyClass(UserWritable.class);
        job.setOutputKeyClass(NullWritable.class);
        //指定job的输入原始文件所在目录
        FileInputFormat.setInputPaths(job, new Path(args[0]));
        //指定job的输出结果所在目录
        FileOutputFormat.setOutputPath(job, new Path(args[1]));
        //将job中配置的相关参数及job所用的java类在的jar包，提交给yarn去运行
        //提交之后，此时客户端代码就执行完毕，退出
        //job.submit();
        //等集群返回结果在退出
        boolean res=job.waitForCompletion(true);
        System.exit(res?0:1);
        //类似于shell中的$？
    }
}

三 通过eclipse将程序打包为mapreduce.jar

四 MapReduce的自定义序列化测试

1 准备数据

[root@centos hadoop-2.7.4]# bin/hdfs dfs -cat /input/data
1    1000    0
2    500    300
1    2000    1000
2    500    200

2 运行MapReduce

[root@centos hadoop-2.7.4]# bin/yarn jar /root/jar/mapreduce.jar /input/data /output3
17/12/20 21:24:45 INFO client.RMProxy: Connecting to ResourceManager at /0.0.0.0:8032
17/12/20 21:24:46 WARN mapreduce.JobResourceUploader: Hadoop command-line option parsing not performed. Implement the Tool interface and execute your application with ToolRunner to remedy this.
17/12/20 21:24:47 INFO input.FileInputFormat: Total input paths to process : 1
17/12/20 21:24:47 INFO mapreduce.JobSubmitter: number of splits:1
17/12/20 21:24:47 INFO mapreduce.JobSubmitter: Submitting tokens for job: job_1513775596077_0001
17/12/20 21:24:49 INFO impl.YarnClientImpl: Submitted application application_1513775596077_0001
17/12/20 21:24:49 INFO mapreduce.Job: The url to track the job: http://centos:8088/proxy/application_1513775596077_0001/
17/12/20 21:24:49 INFO mapreduce.Job: Running job: job_1513775596077_0001
17/12/20 21:25:13 INFO mapreduce.Job: Job job_1513775596077_0001 running in uber mode : false
17/12/20 21:25:13 INFO mapreduce.Job:  map 0% reduce 0%
17/12/20 21:25:38 INFO mapreduce.Job:  map 100% reduce 0%
17/12/20 21:25:54 INFO mapreduce.Job:  map 100% reduce 100%
17/12/20 21:25:56 INFO mapreduce.Job: Job job_1513775596077_0001 completed successfully
17/12/20 21:25:57 INFO mapreduce.Job: Counters: 49
    File System Counters
        FILE: Number of bytes read=94
        FILE: Number of bytes written=241391
        FILE: Number of read operations=0
        FILE: Number of large read operations=0
        FILE: Number of write operations=0
        HDFS: Number of bytes read=135
        HDFS: Number of bytes written=32
        HDFS: Number of read operations=6
        HDFS: Number of large read operations=0
        HDFS: Number of write operations=2
    Job Counters
        Launched map tasks=1
        Launched reduce tasks=1
        Data-local map tasks=1
        Total time spent by all maps in occupied slots (ms)=23672
        Total time spent by all reduces in occupied slots (ms)=11815
        Total time spent by all map tasks (ms)=23672
        Total time spent by all reduce tasks (ms)=11815
        Total vcore-milliseconds taken by all map tasks=23672
        Total vcore-milliseconds taken by all reduce tasks=11815
        Total megabyte-milliseconds taken by all map tasks=24240128
        Total megabyte-milliseconds taken by all reduce tasks=12098560
    Map-Reduce Framework
        Map input records=4
        Map output records=4
        Map output bytes=80
        Map output materialized bytes=94
        Input split bytes=94
        Combine input records=0
        Combine output records=0
        Reduce input groups=2
        Reduce shuffle bytes=94
        Reduce input records=4
        Reduce output records=2
        Spilled Records=8
        Shuffled Maps =1
        Failed Shuffles=0
        Merged Map outputs=1
        GC time elapsed (ms)=157
        CPU time spent (ms)=1090
        Physical memory (bytes) snapshot=275660800
        Virtual memory (bytes) snapshot=4160692224
        Total committed heap usage (bytes)=139264000
    Shuffle Errors
        BAD_ID=0
        CONNECTION=0
        IO_ERROR=0
        WRONG_LENGTH=0
        WRONG_MAP=0
        WRONG_REDUCE=0
    File Input Format Counters
        Bytes Read=41
    File Output Format Counters
        Bytes Written=32

3 测试结果

[root@centos hadoop-2.7.4]# bin/hdfs dfs -cat /output3/part-r-00000
1    3000    1000    2000
2    1000    500    500

五 参考

http://www.jikexueyuan.com/course/2710.html