MapReduce的一对多连接操作

问题描述:

一个trade table表

product1"trade1

product2"trade2

product3"trade3

一个pay table表

product1"pay1

product2"pay2

product2"pay3

product1"pay4

product3"pay5

product3"pay6

建立两个表之间的连接,该两表是一对多关系的

如下:

trade1pay1

trade1pay4

trade2pay2

...

思路:

       为了将两个表整合到一起,由于有相同的第一列,且第一个表与第二个表是一对多关系的。

这里依然采用分组,以及组内排序,只要保证一方最先到达reduce端,则就可以进行迭代处理了。

为了保证第一个表先到达reduce端,可以为定义一个组合键,包含两个值,第一个值为product,第二个值为0或者1,来分别代表第一个表和第二个表,只要按照组内升序排列即可。

具体代码:

自定义组合键策略

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
package whut.onetomany;
import java.io.DataInput;
import java.io.DataOutput;
import java.io.IOException;
import org.apache.hadoop.io.WritableComparable;
public class TextIntPair implements WritableComparable{
    //product1 0/1
    private String firstKey;//product1
    private int secondKey;//0,1;0代表是trade表,1代表是pay表
    //只需要保证trade表在pay表前面就行,则只需要对组顺序排列
                                                           
    public String getFirstKey() {
        return firstKey;
    }
    public void setFirstKey(String firstKey) {
        this.firstKey = firstKey;
    }
    public int getSecondKey() {
        return secondKey;
    }
    public void setSecondKey(int secondKey) {
        this.secondKey = secondKey;
    }
    @Override
    public void write(DataOutput out) throws IOException {
        out.writeUTF(firstKey);
        out.writeInt(secondKey);
    }
    @Override
    public void readFields(DataInput in) throws IOException {
        // TODO Auto-generated method stub
        firstKey=in.readUTF();
        secondKey=in.readInt();
    }
                                                           
    @Override
    public int compareTo(Object o) {
        // TODO Auto-generated method stub
        TextIntPair tip=(TextIntPair)o;
        return this.getFirstKey().compareTo(tip.getFirstKey());
    }
}

分组策略

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
package whut.onetomany;
import org.apache.hadoop.io.WritableComparable;
import org.apache.hadoop.io.WritableComparator;
public class TextComparator extends WritableComparator{
    protected TextComparator() {
        super(TextIntPair.class,true);//注册比较器
    }
    @Override
    public int compare(WritableComparable a, WritableComparable b) {
        // TODO Auto-generated method stub
        TextIntPair tip1=(TextIntPair)a;
        TextIntPair tip2=(TextIntPair)b;
        return tip1.getFirstKey().compareTo(tip2.getFirstKey());
    }
}

组内排序策略:目的是保证第一个表比第二个表先到达

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
package whut.onetomany;
import org.apache.hadoop.io.WritableComparable;
import org.apache.hadoop.io.WritableComparator;
//分组内部进行排序,按照第二个字段进行排序
public class TextIntComparator extends WritableComparator {
    public TextIntComparator()
    {
        super(TextIntPair.class,true);
    }
    //这里可以进行排序的方式管理
    //必须保证是同一个分组的
    //a与b进行比较
    //如果a在前b在后,则会产生升序
    //如果a在后b在前,则会产生降序
    @Override
    public int compare(WritableComparable a, WritableComparable b) {
        // TODO Auto-generated method stub
        TextIntPair ti1=(TextIntPair)a;
        TextIntPair ti2=(TextIntPair)b;
        //首先要保证是同一个组内,同一个组的标识就是第一个字段相同
        if(!ti1.getFirstKey().equals(ti2.getFirstKey()))
           return ti1.getFirstKey().compareTo(ti2.getFirstKey());
        else
           return ti1.getSecondKey()-ti2.getSecondKey();//0,-1,1
    }
                                      
}

分区策略:

1
2
3
4
5
6
7
8
9
10
package whut.onetomany;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Partitioner;
public class PartitionByText extends Partitioner<TextIntPair, Text> {
    @Override
    public int getPartition(TextIntPair key, Text value, int numPartitions) {
        // TODO Auto-generated method stub
        return (key.getFirstKey().hashCode()&Integer.MAX_VALUE)%numPartitions;
    }
}

MapReduce

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
package whut.onetomany;
import java.io.IOException;
import java.util.Iterator;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.conf.Configured;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.Mapper;
import org.apache.hadoop.mapreduce.Mapper.Context;
import org.apache.hadoop.mapreduce.Reducer;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.input.FileSplit;
import org.apache.hadoop.mapreduce.lib.input.MultipleInputs;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
import org.apache.hadoop.util.GenericOptionsParser;
import org.apache.hadoop.util.Tool;
import org.apache.hadoop.util.ToolRunner;
public class JoinMain extends Configured implements Tool {
    public static class JoinMapper extends Mapper<LongWritable, Text, TextIntPair, Text>
    {
        private TextIntPair tp=new TextIntPair();
        private Text val=new Text();
        @Override
        protected void map(LongWritable key, Text value, Context context)
                throws IOException, InterruptedException {
            // TODO Auto-generated method stub
            //获取要处理的文件的名称
            FileSplit file=(FileSplit)context.getInputSplit();
            String fileName=file.getPath().toString();
            //获取输入行分隔
            String line=value.toString();
            String[] lineKeyValue=line.split("\"");
            String lineKey=lineKeyValue[0];
            String lineValue=lineKeyValue[1];
            tp.setFirstKey(lineKey);
            //判断是否是trade文件
            if(fileName.indexOf("trade")>=0)
            {
                tp.setSecondKey(0);
                val.set(lineValue);
            }
            //判断是否是pay文件
            else if(fileName.indexOf("pay")>=0)
            {
                tp.setSecondKey(1);
                val.set(lineValue);
            }
            context.write(tp, val);
        }
    }
                       
    public static class JoinReducer extends Reducer<TextIntPair, Text, Text, Text>
    {
        @Override
        protected void reduce(TextIntPair key, Iterable<Text> values,
                Context context)throws IOException, InterruptedException {
            Iterator<Text> valList=values.iterator();
            //注意这里一定要写成string不可变,写成Text有问题
            //Text trade=valList.next();
            String tradeName=valList.next().toString();
            while(valList.hasNext())
            {
                Text pay=valList.next();
                context.write(new Text(tradeName), pay);
            }
        }
    }
    @Override
    public int run(String[] args) throws Exception
    {
        Configuration conf=getConf();
        Job job=new Job(conf,"JoinJob");
        job.setJarByClass(JoinMain.class);
        //ToolRunner已经利用GenericOptionsParser解析了命令行中的参数
        //并且将其存放在数组中,传递给该run()方法了
        FileInputFormat.addInputPath(job, new Path(args[0]));
        FileInputFormat.addInputPath(job, new Path(args[1]));
        //输入文件必须以,隔开
        //FileInputFormat.addInputPaths(job, args[0]);
        FileOutputFormat.setOutputPath(job, new Path(args[2]));
                           
        job.setMapperClass(JoinMapper.class);
        job.setReducerClass(JoinReducer.class);
        //设置分区方法
        job.setPartitionerClass(PartitionByText.class);
        //设置分组排序
        job.setGroupingComparatorClass(TextComparator.class);
        job.setSortComparatorClass(TextIntComparator.class);
                           
        job.setMapOutputKeyClass(TextIntPair.class);
        job.setMapOutputValueClass(Text.class);
        job.setOutputKeyClass(Text.class);
        job.setOutputValueClass(Text.class);
        job.waitForCompletion(true);
        int exitCode=job.isSuccessful()?0:1;
        return exitCode;
    }
    public static void main(String[] args)throws Exception
    {
        // TODO Auto-generated method stub
        int code=ToolRunner.run(new JoinMain(), args);
        System.exit(code);
    }
}


注意:

     一般有些地方没有定义组内排序策略,但是经过多次测试,发现无法保证第一个表在第二个表之前到达,则这里就自定义了组内排序策略。版本号为Hadoop1.1.2

上一篇:线程安全操作HashMap


下一篇:E-MapReduce中引导操作安装使用Kylin