MapReduce实现分组排序

绝地灬酷狼 2022-06-08 10:52 280阅读 0赞

**MapReduce实现分组排序**

以某次竞赛为例，分别进行如果实现：

*  取每组中男生前三名成绩和女生前三名成绩
 *  按照年龄分组降序输出所有人的成绩
 *  等价的SQL

**0. 预备知识**

**0.1 基于MapReduce实现分组、排序：**

分组： 相当于group by。 MapReduce的实现：相当于分区，以求处理手机上网日志为例，把手机号和非手机号分为两组。

*  在map和reduce阶段进行排序，比较的是k2。v2是不参与排序比较的。如果想让v2也进行排序，需要把k2和v2组装成新的类，作为k2，才能参与比较。
 *  分组时也是按照k2进行比较的。

**0.2 数据准备：**

文件score.txt，并通过hadoop fs -put命令把准备好的数据上传到HDFS上。

jangz   23      male    98
    John    34      male    100
    Tom     45      male    99
    Lily    32      female  40
    Linda   34      female  100
    Chaces  28      male    98
    Dong    29      male    30
    Daniel  33      male    100
    Marvin  24      male    100
    Chaos   30      female  84
    Mei     23      female  90
    Newhire 18      female  100
    Summer  59      male    90

**1. 实现取每组中男生前三名成绩和女生前三名成绩**

**问题分析：**

p1: 取每组中前三名成绩，所以需要进行一次分组，那该如何分组？

p2: 要前三名成绩，则需要对全部数据进行排序，这样才能提取出来前三名，那该如何排序？

p3: 男生和女生？这个又该怎么区分？

**解决思路：**

s1: 在MapReduce中，分组相当于分区，所以我们通过分区的形式实现分组。而分组的根据因素是哪个？当然不可能是成绩来分组，很自然，我们能想到的就是分两组：男生组和女生组。如此，我们直接自定义Partitioner\#getPartition即可。——> 解决p1和p3

s2: 成绩取各组前三的话，如果数据已经排序过，我们直接在各组中取三条记录即可。如果是最简单的操作，也是最节省网络资源的。

**代码实现：**

自定义Bean用于存储数据：

package mapreduce.topk2.top;
    
    import java.io.DataInput;
    import java.io.DataOutput;
    import java.io.IOException;
    
    import org.apache.hadoop.io.WritableComparable;
    
    public class Document implements WritableComparable<Document> {
    
    	private String name;
    
    	private Integer age;
    
    	private String gender;
    
    	private Integer score;
    
    	public Document() {
    	}
    
    	public Document(String name, Integer age, String gender, Integer score) {
    		this.name = name;
    		this.age = age;
    		this.gender = gender;
    		this.score = score;
    	}
    	
    	public void set(String name, Integer age, String gender, Integer score) {
    		this.name = name;
    		this.age = age;
    		this.gender = gender;
    		this.score = score;
    	}
    
    	@Override
    	public void write(DataOutput out) throws IOException {
    		out.writeUTF(name);
    		out.writeInt(age);
    		out.writeUTF(gender);
    		out.writeInt(score);
    	}
    
    	@Override
    	public void readFields(DataInput in) throws IOException {
    		this.name = in.readUTF();
    		this.age = in.readInt();
    		this.gender = in.readUTF();
    		this.score = in.readInt();
    	}
    
    	@Override
    	public int compareTo(Document o) {
    		if (this.score != o.score) {
    			return -this.score.compareTo(o.score);
    		} else {
    			return this.name.compareTo(o.name);
    		}
    	}
    
    	@Override
    	public String toString() {
    		return name + "\t" + age + "\t" + gender + "\t" + score;
    	}
    
    	public String getName() {
    		return name;
    	}
    
    	public void setName(String name) {
    		this.name = name;
    	}
    
    	public Integer getAge() {
    		return age;
    	}
    
    	public void setAge(Integer age) {
    		this.age = age;
    	}
    
    	public String getGender() {
    		return gender;
    	}
    
    	public void setGender(String gender) {
    		this.gender = gender;
    	}
    
    	public Integer getScore() {
    		return score;
    	}
    
    	public void setScore(Integer score) {
    		this.score = score;
    	}
    }

取每组中男生前三名成绩和女生前三名成绩：

package mapreduce.topk2.top;
    
    import java.io.IOException;
    
    import org.apache.hadoop.conf.Configuration;
    import org.apache.hadoop.conf.Configured;
    import org.apache.hadoop.fs.FileSystem;
    import org.apache.hadoop.fs.Path;
    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.Partitioner;
    import org.apache.hadoop.mapreduce.Reducer;
    import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
    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;
    import org.apache.log4j.Logger;
    
    public class Top3GroupByGenderExample extends Configured implements Tool {
    
    	private static final Logger log = Logger.getLogger(Top3GroupByGenderExample.class);
    
    	public static void main(String[] args) throws Exception {
    		Configuration conf = new Configuration();
    		String[] otherArgs = new GenericOptionsParser(conf, args).getRemainingArgs();
    
    		if (otherArgs.length != 2) {
    			log.error("Usage: Top3GroupByGenderExample <in> <out>");
    			System.exit(2);
    		}
    
    		ToolRunner.run(conf, new Top3GroupByGenderExample(), otherArgs);
    	}
    
    	@Override
    	public int run(String[] args) throws Exception {
    		FileSystem fs = FileSystem.get(getConf());
    		Path outPath = new Path(args[1]);
    		if (fs.exists(outPath)) {
    			fs.delete(outPath, true);
    		}
    
    		Job job = Job.getInstance(getConf(), "Top3GroupByGenderExampleJob");
    
    		job.setJarByClass(Top3GroupByGenderExample.class);
    
    		job.setMapperClass(MyMapper.class);
    		job.setMapOutputKeyClass(Document.class);
    		job.setMapOutputValueClass(NullWritable.class);
    		FileInputFormat.setInputPaths(job, new Path(args[0]));
    
    		job.setPartitionerClass(MyPartitioner.class);
    		job.setNumReduceTasks(2);
    
    		job.setReducerClass(MyReducer.class);
    		job.setOutputKeyClass(Document.class);
    		job.setOutputValueClass(NullWritable.class);
    		FileOutputFormat.setOutputPath(job, outPath);
    
    		return job.waitForCompletion(true) ? 0 : 1;
    	}
    
    	public static class MyMapper extends Mapper<LongWritable, Text, Document, NullWritable> {
    
    		private Document document = new Document();
    
    		@Override
    		protected void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException {
    			log.info("MyMapper in<" + key.get() + "," + value.toString() + ">");
    
    			String line = value.toString();
    			String[] infos = line.split("\t");
    
    			String name = infos[0];
    			Integer age = Integer.parseInt(infos[1]);
    			String gender = infos[2];
    			Integer score = Integer.parseInt(infos[3]);
    
    			document.set(name, age, gender, score);
    			context.write(document, NullWritable.get());
    			log.info("MyMapper out<" + document + ">");
    		}
    	}
    
    	public static class MyPartitioner extends Partitioner<Document, NullWritable> {
    
    		@Override
    		public int getPartition(Document key, NullWritable value, int numPartitions) {
    			String gender = key.getGender();
    			return (gender.hashCode() & Integer.MAX_VALUE) % numPartitions;
    		}
    	}
    
    	public static class MyReducer extends Reducer<Document, NullWritable, Document, NullWritable> {
    
    		private int k = 3;
    		private int counter = 0;
    
    		@Override
    		protected void reduce(Document key, Iterable<NullWritable> v2s, Context context)
    				throws IOException, InterruptedException {
    
    			log.info("MyReducer in<" + key + ">");
    
    			if (counter < k) {
    				context.write(key, NullWritable.get());
    				counter += 1;
    
    				log.info("MyReducer out<" + key + ">");
    			}
    		}
    	}
    }

**结果如下：**

执行命令：

![Center][]

得出结果：

![Center 1][]

**2. 按照年龄分组降序输出所有人的成绩**

道理很简单，先按照年龄分组，然后每组成绩降序输出。

代码实现：

自定义Bean：

package mapreduce.topk2.groupsort;
    
    import java.io.DataInput;
    import java.io.DataOutput;
    import java.io.IOException;
    
    import org.apache.hadoop.io.WritableComparable;
    
    public class Person implements WritableComparable<Person> {
    
    	private String name;
    
    	private Integer age;
    
    	private String gender;
    
    	private Integer score;
    
    	public Person() {
    	}
    
    	public Person(String name, Integer age, String gender, Integer score) {
    		this.name = name;
    		this.age = age;
    		this.gender = gender;
    		this.score = score;
    	}
    
    	public void set(String name, Integer age, String gender, Integer score) {
    		this.name = name;
    		this.age = age;
    		this.gender = gender;
    		this.score = score;
    	}
    
    	@Override
    	public String toString() {
    		return name + "\t" + age + "\t" + gender + "\t" + score;
    	}
    
    	@Override
    	public void write(DataOutput out) throws IOException {
    		out.writeUTF(name);
    		out.writeInt(age);
    		out.writeUTF(gender);
    		out.writeInt(score);
    	}
    
    	@Override
    	public void readFields(DataInput in) throws IOException {
    		this.name = in.readUTF();
    		this.age = in.readInt();
    		this.gender = in.readUTF();
    		this.score = in.readInt();
    	}
    
    	/**
    	 * Sort by score desc.
    	 */
    	@Override
    	public int compareTo(Person o) {
    		return -this.score.compareTo(o.score);
    	}
    
    	public String getName() {
    		return name;
    	}
    
    	public Integer getAge() {
    		return age;
    	}
    
    	public String getGender() {
    		return gender;
    	}
    
    	public Integer getScore() {
    		return score;
    	}
    }

实现按年龄分组降序输出所有人的成绩：

package mapreduce.topk2.groupsort;
    
    import java.io.IOException;
    
    import org.apache.hadoop.conf.Configuration;
    import org.apache.hadoop.conf.Configured;
    import org.apache.hadoop.fs.FileSystem;
    import org.apache.hadoop.fs.Path;
    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.Partitioner;
    import org.apache.hadoop.mapreduce.Reducer;
    import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
    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;
    
    /**
     * 
     * <p>Title: GroupByAgeDescScoreExample</p>
     * <p>Description: </p>
     * @author jangz
     * @date 2017/9/29 14:14
     */
    public class GroupByAgeDescScoreExample extends Configured implements Tool {
    
    	public static void main(String[] args) throws Exception {
    		Configuration conf = new Configuration();
    		String[] otherArgs = new GenericOptionsParser(conf, args).getRemainingArgs();
    
    		if (otherArgs.length != 2) {
    			System.out.println("Usage: GroupByAgeDescScoreExample <in> <out>");
    			System.exit(2);
    		}
    
    		ToolRunner.run(conf, new GroupByAgeDescScoreExample(), otherArgs);
    	}
    
    	@Override
    	public int run(String[] args) throws Exception {
    		FileSystem fs = FileSystem.get(getConf());
    		Path outPath = new Path(args[1]);
    		if (fs.exists(outPath)) {
    			fs.delete(outPath, true);
    		}
    
    		Job job = Job.getInstance(getConf(), "GroupByAgeDescScoreExampleJob");
    
    		job.setJarByClass(GroupByAgeDescScoreExample.class);
    
    		job.setMapperClass(MyMapper.class);
    		job.setMapOutputKeyClass(Person.class);
    		job.setMapOutputValueClass(NullWritable.class);
    		FileInputFormat.setInputPaths(job, new Path(args[0]));
    		
    		job.setPartitionerClass(MyPartitioner.class);
    		job.setNumReduceTasks(3);
    		
    		job.setReducerClass(MyReducer.class);
    		job.setOutputKeyClass(Person.class);
    		job.setOutputValueClass(NullWritable.class);
    		FileOutputFormat.setOutputPath(job, outPath);
    
    		return job.waitForCompletion(true) ? 0 : 1;
    	}
    
    	public static class MyMapper extends Mapper<LongWritable, Text, Person, NullWritable> {
    
    		private Person person = new Person();
    
    		@Override
    		protected void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException {
    			System.out.println("MyMapper in<" + key.get() + "," + value.toString() + ">");
    
    			String line = value.toString();
    			String[] infos = line.split("\t");
    
    			String name = infos[0];
    			Integer age = Integer.parseInt(infos[1]);
    			String gender = infos[2];
    			Integer score = Integer.parseInt(infos[3]);
    
    			person.set(name, age, gender, score);
    			context.write(person, NullWritable.get());
    			System.out.println("MyMapper out<" + person + ">");
    		}
    	}
    	
    	public static class MyPartitioner extends Partitioner<Person, NullWritable> {
    
    		@Override
    		public int getPartition(Person key, NullWritable value, int numPartitions) {
    			
    			Integer age = key.getAge();
    			
    			if (age < 20) {
    				return 0;
    			} else if (age <= 50) {
    				return 1;
    			} else {
    				return 2;
    			}
    		}
    	}
    
    	public static class MyReducer extends Reducer<Person, NullWritable, Person, NullWritable> {
    
    		private Text k = new Text();
    
    		@Override
    		protected void reduce(Person key, Iterable<NullWritable> v2s, Context context)
    				throws IOException, InterruptedException {
    			System.out.println("MyReducer in<" + key + ">");
    
    			context.write(key, NullWritable.get());
    
    			System.out.println("MyReducer out<" + k + "," + key + ">");
    		}
    	}
    }

**结果如下：**

执行命令：

![Center 2][]

得出结果：

![Center 3][]

**3. 等价的SQL**

SQL脚本：

CREATE TABLE score (
    	id INT PRIMARY KEY AUTO_INCREMENT,
    	name VARCHAR(50),
    	age INT,
    	gender VARCHAR(10),
    	score INT
    );
    
    INSERT INTO score(name, age, gender, score)
    VALUES('jangz', 23, 'male', 98),
    ('John', 34, 'male', 100),
    ('Tom', 45, 'male', 99),
    ('Lily', 32, 'female', 40),
    ('Linda', 34, 'female', 100),
    ('Chaces', 28, 'male', 98),
    ('Dong', 29, 'male', 30),
    ('Daniel', 33, 'male', 100),
    ('Marvin', 24, 'male', 100),
    ('Chaos', 30, 'female', 84),
    ('Mei', 23, 'female', 90),
    ('Newhire', 18, 'female', 100),
    ('Summer', 59, 'male', 90);

3.1 取每组中男生前三名成绩和女生前三名成绩

采用MapReduce的‘分而治之’的思想：

(SELECT name, age, gender, score
    FROM score
    WHERE gender='female'
    ORDER BY gender, score DESC, name
    LIMIT 3)
    UNION ALL
    (SELECT name, age, gender, score
    FROM score
    WHERE gender='male'
    ORDER BY gender, score DESC, name
    LIMIT 3);

3.2 按照年龄分组降序输出所有人的成绩

SELECT name, age, gender, score
    FROM score
    GROUP BY age
    ORDER BY score DESC, name

**Summary**

1. 在map和reduce阶段进行排序，比较的是k2，v2不参与排序比较。（map和reduce阶段都有partition、sort、combine操作）。

2. 为了数据操作方便，我们可以自定义Bean，并让其实现WritableComparable接口，重写write、readFields和compareTo方法，当然，一定要记得重写toString方法，不然reduce最终输出的结果会达不到预期。

3. 针对非同类数据进行分组，比如按照年龄段，那么我们可以采用分区的形式实现分组。

但是针对同类数据进行分组，比如ip（长度是固定的）和出现次数组合键k2，我们可以采用GroupingComparator。

4. MapReduce是什么？MapReduce是一个运行在大规模集群上，能够可靠且容错地并行处理海量数据集的软件框架。

针对代码，大家也可以到[博主的GitHub仓库][GitHub]进行下载，自己打包运行。博主设计，仅供参考！

[Center]: /images/20220608/e5606a5015314dbd82cbd532e130dcd9.png
[Center 1]: /images/20220608/1dce5302395243e793d5f7f7445d3e8b.png
[Center 2]: /images/20220608/eb050335612242aa915f1304aebd9a1d.png
[Center 3]: /images/20220608/8c79e872ab8d4563a556f97c61474dd5.png
[GitHub]: https://github.com/Zychaowill/HadoopLearning/tree/master/src/main/java/mapreduce/topk2