Hadoop中Join多种应用

总结:

Join: 将两个中的字段,通过公共字段进行关联! 
        MR :
                ①在Reduce端Join
                        1.保证两个文件中的所有字段,都必须达到Reduce!需要将两个文件的所有字段封装为一个Bean
                        2.在Map端,为每个文件,打上标记
                        
                        弊端:如果数据量大,reduce端处理过程耗时!               
                ②在Map端Join
                
                        1.将大文件,作为输入文件,通过输入格式读入到MapTask
                        2.小文件,使用分布式缓存,在进入Mapper时,提前从缓存中读取小文件中的内容
                        
                            缓存文件:  job.addCacheFile()
                            读取缓存:  jod.getCacheFiles()
                        

1 Reduce Join 

2 Reduce Join案例实操

1.需求

表订单数据表t_order

id

pid

amount

1001

01

1

1002

02

2

1003

03

3

1004

01

4

1005

02

5

1006

03

6

表商品信息表t_product

pid

pname

01

小米

02

华为

03

格力

       将商品信息表中数据根据商品pid合并到订单数据表中。

表最终数据形式

id

pname

amount

1001

小米

1

1004

小米

4

1002

华为

2

1005

华为

5

1003

格力

3

1006

格力

6

2.需求分析

通过将关联条件作为Map输出的key,将两表满足Join条件的数据并携带数据所来源的文件信息,发往同一个ReduceTask,在Reduce中进行数据的串联,如图所示。

                                                                                  图 Reduce端表合并

3.代码实现

1)创建商品和订合并后的Bean类

package com.demo.mapreduce.table;

import java.io.DataInput;

import java.io.DataOutput;

import java.io.IOException;

import org.apache.hadoop.io.Writable;

 

public class TableBean implements Writable {

 

    private String order_id; // 订单id

    private String p_id;      // 产品id

    private int amount;       // 产品数量

    private String pname;     // 产品名称

    private String flag;      // 表的标记

 

    public TableBean() {

       super();

    }

 

    public TableBean(String order_id, String p_id, int amount, String pname, String flag) {

 

       super();

 

       this.order_id = order_id;

       this.p_id = p_id;

       this.amount = amount;

       this.pname = pname;

       this.flag = flag;

    }

 

    public String getFlag() {

       return flag;

    }

 

    public void setFlag(String flag) {

       this.flag = flag;

    }

 

    public String getOrder_id() {

       return order_id;

    }

 

    public void setOrder_id(String order_id) {

       this.order_id = order_id;

    }

 

    public String getP_id() {

       return p_id;

    }

 

    public void setP_id(String p_id) {

       this.p_id = p_id;

    }

 

    public int getAmount() {

       return amount;

    }

 

    public void setAmount(int amount) {

       this.amount = amount;

    }

 

    public String getPname() {

       return pname;

    }

 

    public void setPname(String pname) {

       this.pname = pname;

    }

 

    @Override

    public void write(DataOutput out) throws IOException {

       out.writeUTF(order_id);

       out.writeUTF(p_id);

       out.writeInt(amount);

       out.writeUTF(pname);

       out.writeUTF(flag);

    }

 

    @Override

    public void readFields(DataInput in) throws IOException {

       this.order_id = in.readUTF();

       this.p_id = in.readUTF();

       this.amount = in.readInt();

       this.pname = in.readUTF();

       this.flag = in.readUTF();

    }

 

    @Override

    public String toString() {

       return order_id + "\t" + pname + "\t" + amount + "\t" ;

    }

}

2)编写TableMapper类

package com.demo.mapreduce.table;

import java.io.IOException;

import org.apache.hadoop.io.LongWritable;

import org.apache.hadoop.io.Text;

import org.apache.hadoop.mapreduce.Mapper;

import org.apache.hadoop.mapreduce.lib.input.FileSplit;

 

public class TableMapper extends Mapper<LongWritable, Text, Text, TableBean>{

 

String name;

    TableBean bean = new TableBean();

    Text k = new Text();

   

    @Override

    protected void setup(Context context) throws IOException, InterruptedException {

 

       // 1 获取输入文件切片

       FileSplit split = (FileSplit) context.getInputSplit();

 

       // 2 获取输入文件名称

       name = split.getPath().getName();

    }

 

    @Override

    protected void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException {

      

       // 1 获取输入数据

       String line = value.toString();

      

       // 2 不同文件分别处理

       if (name.startsWith("order")) {// 订单表处理

 

           // 2.1 切割

           String[] fields = line.split("\t");

          

           // 2.2 封装bean对象

           bean.setOrder_id(fields[0]);

           bean.setP_id(fields[1]);

           bean.setAmount(Integer.parseInt(fields[2]));

           bean.setPname("");

           bean.setFlag("order");

          

           k.set(fields[1]);

       }else {// 产品表处理

 

           // 2.3 切割

           String[] fields = line.split("\t");

          

           // 2.4 封装bean对象

           bean.setP_id(fields[0]);

           bean.setPname(fields[1]);

           bean.setFlag("pd");

           bean.setAmount(0);

           bean.setOrder_id("");

          

           k.set(fields[0]);

       }

 

       // 3 写出

       context.write(k, bean);

    }

}

3)编写TableReducer类

package com.demo.mapreduce.table;

import java.io.IOException;

import java.util.ArrayList;

import org.apache.commons.beanutils.BeanUtils;

import org.apache.hadoop.io.NullWritable;

import org.apache.hadoop.io.Text;

import org.apache.hadoop.mapreduce.Reducer;

 

public class TableReducer extends Reducer<Text, TableBean, TableBean, NullWritable> {

 

    @Override

    protected void reduce(Text key, Iterable<TableBean> values, Context context)  throws IOException, InterruptedException {

 

       // 1准备存储订单的集合

       ArrayList<TableBean> orderBeans = new ArrayList<>();

      

// 2 准备bean对象

       TableBean pdBean = new TableBean();

 

       for (TableBean bean : values) {

 

           if ("order".equals(bean.getFlag())) {// 订单表

 

              // 拷贝传递过来的每条订单数据到集合中

              TableBean orderBean = new TableBean();

 

              try {

                  BeanUtils.copyProperties(orderBean, bean);

              } catch (Exception e) {

                  e.printStackTrace();

              }

 

              orderBeans.add(orderBean);

           } else {// 产品表

 

              try {

                  // 拷贝传递过来的产品表到内存中

                  BeanUtils.copyProperties(pdBean, bean);

              } catch (Exception e) {

                  e.printStackTrace();

              }

           }

       }

 

       // 3 表的拼接

       for(TableBean bean:orderBeans){

 

           bean.setPname (pdBean.getPname());

          

           // 4 数据写出去

           context.write(bean, NullWritable.get());

       }

    }

}

4)编写TableDriver类

package com.demo.mapreduce.table;

import org.apache.hadoop.conf.Configuration;

import org.apache.hadoop.fs.Path;

import org.apache.hadoop.io.NullWritable;

import org.apache.hadoop.io.Text;

import org.apache.hadoop.mapreduce.Job;

import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;

import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;

 

public class TableDriver {

 

    public static void main(String[] args) throws Exception {

      

// 0 根据自己电脑路径重新配置

args = new String[]{"e:/input/inputtable","e:/output1"};

 

// 1 获取配置信息,或者job对象实例

       Configuration configuration = new Configuration();

       Job job = Job.getInstance(configuration);

 

       // 2 指定本程序的jar包所在的本地路径

       job.setJarByClass(TableDriver.class);

 

       // 3 指定本业务job要使用的Mapper/Reducer业务类

       job.setMapperClass(TableMapper.class);

       job.setReducerClass(TableReducer.class);

 

       // 4 指定Mapper输出数据的kv类型

       job.setMapOutputKeyClass(Text.class);

       job.setMapOutputValueClass(TableBean.class);

 

       // 5 指定最终输出的数据的kv类型

       job.setOutputKeyClass(TableBean.class);

       job.setOutputValueClass(NullWritable.class);

 

       // 6 指定job的输入原始文件所在目录

       FileInputFormat.setInputPaths(job, new Path(args[0]));

       FileOutputFormat.setOutputPath(job, new Path(args[1]));

 

       // 7 将job中配置的相关参数,以及job所用的java类所在的jar包, 提交给yarn去运行

       boolean result = job.waitForCompletion(true);

       System.exit(result ? 0 : 1);

    }

}

4.测试

运行程序查看结果

1001   小米    1  

1001   小米    1  

1002   华为    2  

1002   华为    2  

1003   格力    3  

1003   格力    3    

5.总结

 

3 Map Join

1.使用场景

Map Join适用于一张表十分小、一张表很大的场景。

2.优点

思考:在Reduce端处理过多的表,非常容易产生数据倾斜。怎么办?

Map端缓存多张表,提前处理业务逻辑,这样增加Map端业务,减少Reduce端数据的压力,尽可能的减少数据倾斜。

3.具体办法:采用DistributedCache

       (1)在Mapper的setup阶段,将文件读取到缓存集合中。

       (2)在驱动函数中加载缓存

// 缓存普通文件到Task运行节点。

job.addCacheFile(new URI("file://e:/cache/pd.txt"));

4 Map Join案例实操

表4-4 订单数据表t_order

id

pid

amount

1001

01

1

1002

02

2

1003

03

3

1004

01

4

1005

02

5

1006

03

6

表4-5 商品信息表t_product

pid

pname

01

小米

02

华为

03

格力

       将商品信息表中数据根据商品pid合并到订单数据表中。

表4-6 最终数据形式

id

pname

amount

1001

小米

1

1004

小米

4

1002

华为

2

1005

华为

5

1003

格力

3

1006

格力

6

2.需求分析

MapJoin适用于关联表中有小表的情形

3.实现代码

(1)先在驱动模块中添加缓存文件

package test;

import java.net.URI;

import org.apache.hadoop.conf.Configuration;

import org.apache.hadoop.fs.Path;

import org.apache.hadoop.io.NullWritable;

import org.apache.hadoop.io.Text;

import org.apache.hadoop.mapreduce.Job;

import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;

import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;

 

public class DistributedCacheDriver {

 

    public static void main(String[] args) throws Exception {

      

// 0 根据自己电脑路径重新配置

args = new String[]{"e:/input/inputtable2", "e:/output1"};

 

// 1 获取job信息

       Configuration configuration = new Configuration();

       Job job = Job.getInstance(configuration);

 

       // 2 设置加载jar包路径

       job.setJarByClass(DistributedCacheDriver.class);

 

       // 3 关联map

       job.setMapperClass(DistributedCacheMapper.class);

      

// 4 设置最终输出数据类型

       job.setOutputKeyClass(Text.class);

       job.setOutputValueClass(NullWritable.class);

 

       // 5 设置输入输出路径

       FileInputFormat.setInputPaths(job, new Path(args[0]));

       FileOutputFormat.setOutputPath(job, new Path(args[1]));

 

       // 6 加载缓存数据

       job.addCacheFile(new URI("file:///e:/input/inputcache/pd.txt"));

      

       // 7 MapJoin的逻辑不需要Reduce阶段,设置reduceTask数量为0

       job.setNumReduceTasks(0);

 

       // 8 提交

       boolean result = job.waitForCompletion(true);

       System.exit(result ? 0 : 1);

    }

}

(2)读取缓存的文件数据

package test;

import java.io.BufferedReader;

import java.io.FileInputStream;

import java.io.IOException;

import java.io.InputStreamReader;

import java.util.HashMap;

import java.util.Map;

import org.apache.commons.lang.StringUtils;

import org.apache.hadoop.io.LongWritable;

import org.apache.hadoop.io.NullWritable;

import org.apache.hadoop.io.Text;

import org.apache.hadoop.mapreduce.Mapper;

 

public class DistributedCacheMapper extends Mapper<LongWritable, Text, Text, NullWritable>{

 

    Map<String, String> pdMap = new HashMap<>();

   

    @Override

    protected void setup(Mapper<LongWritable, Text, Text, NullWritable>.Context context) throws IOException, InterruptedException {

 

       // 1 获取缓存的文件

       URI[] cacheFiles = context.getCacheFiles();

       String path = cacheFiles[0].getPath().toString();

      

       BufferedReader reader = new BufferedReader(new InputStreamReader(new FileInputStream(path), "UTF-8"));

      

       String line;

       while(StringUtils.isNotEmpty(line = reader.readLine())){

 

           // 2 切割

           String[] fields = line.split("\t");

          

           // 3 缓存数据到集合

           pdMap.put(fields[0], fields[1]);

       }

      

       // 4 关流

       reader.close();

    }

   

    Text k = new Text();

   

    @Override

    protected void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException {

 

       // 1 获取一行

       String line = value.toString();

      

       // 2 截取

       String[] fields = line.split("\t");

      

       // 3 获取产品id

       String pId = fields[1];

      

       // 4 获取商品名称

       String pdName = pdMap.get(pId);

      

       // 5 拼接

       k.set(line + "\t"+ pdName);

      

       // 6 写出

       context.write(k, NullWritable.get());

    }

}

  • 5
    点赞
  • 24
    收藏
    觉得还不错? 一键收藏
  • 4
    评论
评论 4
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值