MapReduce实战之 MapReduce中多表合并案例

 MapReduce中多表合并案例

1)需求:

订单数据表t_order

id

pid

amount

1001

01

1

1002

02

2

1003

03

3

1001    01    1
1002    02    2
1003    03    3
1004    01    4
1005    02    5
1006    03    6

商品信息表t_product

pid

pname

01

小米

02

华为

03

格力

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。

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

 

 需求1:Reduce端表合并(数据倾斜)

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reduce task,在reduce中进行数据的串联。

1)创建商品和订合并后的bean

package com.atguigu.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程序

package com.atguigu.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>{

       TableBean bean = new TableBean();

       Text k = new Text();

      

       @Override

       protected void map(LongWritable key, Text value, Context context)

                     throws IOException, InterruptedException {

             

              // 1 获取输入文件类型

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

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

             

              // 2 获取输入数据

              String line = value.toString();

             

              // 3 不同文件分别处理

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

                     // 3.1 切割

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

                    

                     // 3.2 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("0");

                    

                     k.set(fields[1]);

              }else {// 产品表处理

                     // 3.3 切割

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

                    

                     // 3.4 封装bean对象

                     bean.setP_id(fields[0]);

                     bean.setPname(fields[1]);

                     bean.setFlag("1");

                     bean.setAmount(0);

                     bean.setOrder_id("");

                    

                     k.set(fields[0]);

              }

              // 4 写出

              context.write(k, bean);

       }

}

3)编写TableReducer程序

package com.atguigu.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 ("0".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程序

package com.atguigu.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 {

              // 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 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);

       }

}

3)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    

缺点:这种方式中,合并的操作是在reduce阶段完成,reduce端的处理压力太大,map节点的运算负载则很低,资源利用率不高,且在reduce阶段极易产生数据倾斜

解决方案: map端实现数据合并

 

需求2:Map端表合并

1)分析

适用于关联表中有小表的情形;

可以将小表分发到所有的map节点,这样,map节点就可以在本地对自己所读到的大表数据进行合并并输出最终结果,可以大大提高合并操作的并发度,加快处理速度。

 

2)实操案例

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 {

              // 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:/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

              BufferedReader reader = new BufferedReader(new InputStreamReader(new FileInputStream("pd.txt"),"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());

       }

}

 

评论
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