统计和TopKey

key和value的默认分隔符为tab键

设置分隔符


程序一

package org.conan.myhadoop.TopKey;

import java.io.IOException;

import org.apache.hadoop.conf.Configuration;
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.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.input.TextInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;

//单文件最值
public class TopKMapReduce {

    static class TopKMapper extends
            Mapper<LongWritable, Text, Text, LongWritable> {
        // 输出的key
        private Text mapOutputKey = new Text();
        // 输出的value
        private LongWritable mapOutputValue = new LongWritable();

        // 存储最大值和初始值
        long topkValue = Long.MIN_VALUE;

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

            String lineValue = value.toString();

            String[] strs = lineValue.split("\t");
            // 中间值
            long tempValue = Long.valueOf(strs[1]);

            if (topkValue < tempValue) {

                topkValue = tempValue;
                mapOutputKey.set(strs[0]);
            }
        }

        @Override
        protected void cleanup(Context context) throws IOException,
                InterruptedException {
            mapOutputValue.set(topkValue);
            context.write(mapOutputKey, mapOutputValue);
        }

        @Override
        protected void setup(Context context) throws IOException,
                InterruptedException {
            super.setup(context);
        }

    }

    public int run(String[] args) throws Exception {

        Configuration conf = new Configuration();
        Job job = new Job(conf, TopKMapReduce.class.getSimpleName());
        job.setJarByClass(TopKMapReduce.class);
        Path inputDir = new Path(args[0]);
        FileInputFormat.addInputPath(job, inputDir);
        job.setInputFormatClass(TextInputFormat.class);
        job.setMapperClass(TopKMapper.class);

        job.setMapOutputKeyClass(Text.class);
        job.setMapOutputValueClass(LongWritable.class);
        // job.setReducerClass(ModuleReducer.class);
        // job.setOutputKeyClass(LongWritable.class);
        // job.setOutputValueClass(Text.class);
        job.setNumReduceTasks(0);

        Path outputDir = new Path(args[1]);
        FileOutputFormat.setOutputPath(job, outputDir);
        Boolean isCompletion = job.waitForCompletion(true);
        return isCompletion ? 0 : 1;
    }

    public static void main(String[] args) throws Exception {
        args = new String[] { "hdfs://hadoop-master:9000/data/wcoutput",
                "hdfs://hadoop-master:9000/data/topkoutput" };
        int status = new TopKMapReduce().run(args);
        System.exit(status);
    }

}


程序二

package org.conan.myhadoop.TopKey;

import java.io.IOException;
import java.util.Set;
import java.util.TreeMap;

import org.apache.hadoop.conf.Configuration;
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.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.input.TextInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;

//单文件 top n TreeMap实现
public class TopKMapReduceV2 {

    static class TopKMapper extends
            Mapper<LongWritable, Text, Text, LongWritable> {
        
        public static final int K=3;//前三名
        private LongWritable mapKey = new LongWritable();
        private Text mapValue = new Text();
        
        TreeMap<LongWritable, Text> topMap = null;//默认按key的升序排列

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

            String lineValue = value.toString();

            String[] strs = lineValue.split("\t");
        
            long tempValue = Long.valueOf(strs[1]);
               String tempKey=strs[0];
               mapKey.set(tempValue);
               mapValue.set(tempKey);
               topMap.put(mapKey, mapValue);
    
               if(topMap.size()>K){
                   topMap.remove(topMap.firstKey());
                   
               }
        }

        @Override
        protected void cleanup(Context context) throws IOException,
                InterruptedException {
            Set<LongWritable> keySet=    topMap.keySet();
            for( LongWritable key:keySet) {
                
                context.write(topMap.get(key), key);
            }
        }

        @Override
        protected void setup(Context context) throws IOException,
                InterruptedException {
            super.setup(context);
        }

    }

    public int run(String[] args) throws Exception {

        Configuration conf = new Configuration();
        Job job = new Job(conf, TopKMapReduceV2.class.getSimpleName());
        job.setJarByClass(TopKMapReduceV2.class);
        Path inputDir = new Path(args[0]);
        FileInputFormat.addInputPath(job, inputDir);
        job.setInputFormatClass(TextInputFormat.class);
        job.setMapperClass(TopKMapper.class);

        job.setMapOutputKeyClass(Text.class);
        job.setMapOutputValueClass(LongWritable.class);
        // job.setReducerClass(ModuleReducer.class);
        // job.setOutputKeyClass(LongWritable.class);
        // job.setOutputValueClass(Text.class);
        job.setNumReduceTasks(0);

        Path outputDir = new Path(args[1]);
        FileOutputFormat.setOutputPath(job, outputDir);
        Boolean isCompletion = job.waitForCompletion(true);
        return isCompletion ? 0 : 1;
    }

    public static void main(String[] args) throws Exception {
        args = new String[] { "hdfs://hadoop-master:9000/data/wcoutput",
                "hdfs://hadoop-master:9000/data/topkoutput2" };
        int status = new TopKMapReduceV2().run(args);
        System.exit(status);
    }

}


程序三

package org.conan.myhadoop.TopKey;

import java.io.IOException;
import java.util.Comparator;

import java.util.TreeSet;

import org.apache.hadoop.conf.Configuration;
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.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.input.TextInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
//单文件 top n TreeSet实现
public class TopKMapReduceV3 {

    static class TopKMapper extends
            Mapper<LongWritable, Text, Text, LongWritable> {
        
        public static final int K=3;//前三名
        
        
        TreeSet<TopKWritable> topSet = new TreeSet<TopKWritable>(//
                new Comparator<TopKWritable>() {

            @Override
            public int compare(TopKWritable o1, TopKWritable o2) {
            
                return o1.getCount().compareTo(o2.getCount());
            }
        }) ;

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

            String lineValue = value.toString();

            String[] strs = lineValue.split("\t");
        
            long tempValue = Long.valueOf(strs[1]);
             
              
               topSet.add(new TopKWritable(strs[0], tempValue));
    
               if(topSet.size()>K){
                   topSet.remove(topSet.first());
                   
               }
        }

        @Override
        protected void cleanup(Context context) throws IOException,
                InterruptedException {
            
            for( TopKWritable top:topSet) {
                
                context.write(new Text(top.getWord()), new LongWritable(top.getCount()));
            }
        }

        @Override
        protected void setup(Context context) throws IOException,
                InterruptedException {
            super.setup(context);
        }

    }

    public int run(String[] args) throws Exception {

        Configuration conf = new Configuration();
        Job job = new Job(conf, TopKMapReduceV3.class.getSimpleName());
        job.setJarByClass(TopKMapReduceV3.class);
        Path inputDir = new Path(args[0]);
        FileInputFormat.addInputPath(job, inputDir);
        job.setInputFormatClass(TextInputFormat.class);
        job.setMapperClass(TopKMapper.class);

        job.setMapOutputKeyClass(Text.class);
        job.setMapOutputValueClass(LongWritable.class);
        // job.setReducerClass(ModuleReducer.class);
        // job.setOutputKeyClass(LongWritable.class);
        // job.setOutputValueClass(Text.class);
        job.setNumReduceTasks(0);

        Path outputDir = new Path(args[1]);
        FileOutputFormat.setOutputPath(job, outputDir);
        Boolean isCompletion = job.waitForCompletion(true);
        return isCompletion ? 0 : 1;
    }

    public static void main(String[] args) throws Exception {
        args = new String[] { "hdfs://hadoop-master:9000/data/wcoutput",
                "hdfs://hadoop-master:9000/data/topkoutput3" };
        int status = new TopKMapReduceV3().run(args);
        System.exit(status);
    }

}


程序四  自定义数据类型加比较器

package org.conan.myhadoop.TopKey;

import java.io.IOException;
import java.util.Comparator;

import java.util.TreeSet;

import org.apache.hadoop.conf.Configuration;
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.Reducer;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.input.TextInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;

//多个文件,需要reduce统计top n
public class TopKMapReduceV4 {

    static class TopKMapper extends
            Mapper<LongWritable, Text, Text, LongWritable> {

        @Override
        public void map(LongWritable key, Text value, Context context)
                throws IOException, InterruptedException {

            String lineValue = value.toString();

            String[] strs = lineValue.split("\t");

            long tempValue = Long.valueOf(strs[1]);

            context.write(new Text(strs[0]), new LongWritable(tempValue));
        }

        @Override
        public void cleanup(Context context) throws IOException,
                InterruptedException {

            super.cleanup(context);
        }

        @Override
        public void setup(Context context) throws IOException,
                InterruptedException {
            super.setup(context);
        }

    }

    public static class TopKReducer extends
            Reducer<Text, LongWritable, Text, LongWritable> {
        public static final int K = 3;// 前三名
        TreeSet<TopKWritable> topSet = new TreeSet<TopKWritable>(//
                new Comparator<TopKWritable>() {

                    @Override
                    public int compare(TopKWritable o1, TopKWritable o2) {

                        return o1.getCount().compareTo(o2.getCount());
                    }
                });

        @Override
        public void setup(Context context) throws IOException,
                InterruptedException {

            super.setup(context);
        }

        @Override
        public void reduce(Text key, Iterable<LongWritable> values,
                Context context) throws IOException, InterruptedException {
            long count = 0;
            for (LongWritable value : values) {

                count += value.get();
            }

            topSet.add(new TopKWritable(key.toString(), count));
            if (topSet.size() > K) {

                topSet.remove(topSet.first());
            }
        }

        @Override
        public void cleanup(Context context) throws IOException,
                InterruptedException {
            
            for (TopKWritable top : topSet) {
                context.write(new Text(top.getWord()),
                        new LongWritable(top.getCount()));
            }
        }

    }

    public int run(String[] args) throws Exception {

        Configuration conf = new Configuration();
        Job job = new Job(conf, TopKMapReduceV4.class.getSimpleName());
        job.setJarByClass(TopKMapReduceV4.class);
        Path inputDir = new Path(args[0]);
        FileInputFormat.addInputPath(job, inputDir);
        job.setInputFormatClass(TextInputFormat.class);
        job.setMapperClass(TopKMapper.class);

        job.setMapOutputKeyClass(Text.class);
        job.setMapOutputValueClass(LongWritable.class);
        job.setReducerClass(TopKReducer.class);
         job.setOutputKeyClass(Text.class);
         job.setOutputValueClass(LongWritable.class);
        job.setNumReduceTasks(1);

        Path outputDir = new Path(args[1]);
        FileOutputFormat.setOutputPath(job, outputDir);
        Boolean isCompletion = job.waitForCompletion(true);
        return isCompletion ? 0 : 1;
    }

    public static void main(String[] args) throws Exception {
        args = new String[] { "hdfs://hadoop-master:9000/data/wcoutput",
                "hdfs://hadoop-master:9000/data/topkoutput4" };
        int status = new TopKMapReduceV4().run(args);
        System.exit(status);
    }

}
package org.conan.myhadoop.TopKey;

import java.io.DataInput;
import java.io.DataOutput;
import java.io.IOException;

import org.apache.hadoop.io.WritableComparable;

//自定义数据类型
public class TopKWritable implements WritableComparable<TopKWritable> {

    private String word;
    private Long count;
    public TopKWritable(){};
    public TopKWritable(String word,Long count) {
        this.set(word, count);
    }
    public void set(String word,Long count) {
        this.word = word;
        this.count = count;
    }
    public String getWord() {
        return word;
    }
    public Long getCount() {
        return count;
    }



    @Override
    public void write(DataOutput out) throws IOException {
        
        out.writeUTF(word);
        out.writeLong(count);
    }
    @Override
    public void readFields(DataInput in) throws IOException {
    
        this.word=in.readUTF();
        this.count=in.readLong();
    }

    

    @Override
    public int compareTo(TopKWritable o) {
        int cmp=this.word.compareTo(o.getWord());
        if(0!=cmp){
            
            return cmp;
        }
        
        return this.count.compareTo(o.getCount());
    }
    
    
    
    @Override
    public String toString() {
        return word +"\t"+count;
    }
    
    @Override
    public int hashCode() {
        final int prime = 31;
        int result = 1;
        result = prime * result + ((count == null) ? 0 : count.hashCode());
        result = prime * result + ((word == null) ? 0 : word.hashCode());
        return result;
    }
    @Override
    public boolean equals(Object obj) {
        if (this == obj)
            return true;
        if (obj == null)
            return false;
        if (getClass() != obj.getClass())
            return false;
        TopKWritable other = (TopKWritable) obj;
        if (count == null) {
            if (other.count != null)
                return false;
        } else if (!count.equals(other.count))
            return false;
        if (word == null) {
            if (other.word != null)
                return false;
        } else if (!word.equals(other.word))
            return false;
        return true;
    }
    
    
    

    
    
    
}


程序五:经典案例


wKioL1ZJsWvzQRckAABJRXNZp1E945.png



package org.conan.myhadoop.TopKey;

import java.io.IOException;
import java.util.TreeSet;

import org.apache.hadoop.conf.Configuration;
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.Reducer;

import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.input.TextInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;

/**
 * 
 数据格式: 语言类别 歌曲名称 收藏次数 播放次数 歌手名称
 * 
 * 需求: 统计前十首播放最多的歌曲名称和次数
 * 
 * 
 */
public class TopKeyMapReduce {
    public static final int K = 10;

    static class TopKeyMapper extends
            Mapper<LongWritable, Text, Text, LongWritable> {

        @Override
        public void map(LongWritable key, Text value, Context context)
                throws IOException, InterruptedException {

            String lineValue = value.toString();
            if (null == lineValue) {
                return;
            }
            String[] strs = lineValue.split("\t");
           if (null!=strs&&strs.length==5){
               
               String languageType=strs[0];
               String singName=strs[1];
               String playTimes=strs[3];
               context.write(//
                       new Text(languageType+"\t"+ singName),//
                       new LongWritable(Long.valueOf(playTimes)));
           }
            

            
        }

        @Override
        public void cleanup(Context context) throws IOException,
                InterruptedException {

            super.cleanup(context);
        }

        @Override
        public void setup(Context context) throws IOException,
                InterruptedException {
            super.setup(context);
        }

    }

    public static class TopKeyReducer extends
            Reducer<Text, LongWritable, TopKeyWritable, NullWritable> {

        TreeSet<TopKeyWritable> topSet = new TreeSet<TopKeyWritable>();

        @Override
        public void setup(Context context) throws IOException,
                InterruptedException {

            super.setup(context);
        }

        @Override
        public void reduce(Text key, Iterable<LongWritable> values,
                Context context) throws IOException, InterruptedException {
            if (null==key){
                return;
            }
            
            String[] splited =key.toString().split("\t");
            if(null==splited||splited.length==0){
                return ;
            }
            
            String languageType=splited[0];
            String singName=splited[1];
            
            
            Long playTimes=0L; 
            for (LongWritable value : values) {

                playTimes += value.get();
            }

            topSet.add(new TopKeyWritable(languageType, singName, playTimes));
        
            if (topSet.size() > K) {

                topSet.remove(topSet.last());
            }
        }

        @Override
        public void cleanup(Context context) throws IOException,
                InterruptedException {

            for (TopKeyWritable top : topSet) {
                context.write(top,NullWritable.get());
            }
        }

    }

    public int run(String[] args) throws Exception {

        Configuration conf = new Configuration();
        Job job = new Job(conf, TopKeyMapReduce.class.getSimpleName());
        job.setJarByClass(TopKeyMapReduce.class);
        Path inputDir = new Path(args[0]);
        FileInputFormat.addInputPath(job, inputDir);
        job.setInputFormatClass(TextInputFormat.class);
        
        job.setMapperClass(TopKeyMapper.class);
        job.setMapOutputKeyClass(Text.class);
        job.setMapOutputValueClass(LongWritable.class);
        
        job.setReducerClass(TopKeyReducer.class);
        job.setOutputKeyClass(TopKeyWritable.class);
        job.setOutputValueClass(NullWritable.class);
        job.setNumReduceTasks(1);

        Path outputDir = new Path(args[1]);
        FileOutputFormat.setOutputPath(job, outputDir);
        Boolean isCompletion = job.waitForCompletion(true);
        return isCompletion ? 0 : 1;
    }

    public static void main(String[] args) throws Exception {
        args = new String[] { "hdfs://hadoop-master:9000/data/topkey/input",
                "hdfs://hadoop-master:9000/data/topkey/output" };
        int status = new TopKMapReduceV4().run(args);
        System.exit(status);
    }
}
package org.conan.myhadoop.TopKey;

import java.io.DataInput;
import java.io.DataOutput;
import java.io.IOException;

import org.apache.hadoop.io.WritableComparable;

public class TopKeyWritable implements WritableComparable<TopKeyWritable> {
    String languageType;
    String singName;
    Long playTimes;

    public TopKeyWritable() {
    };

    public TopKeyWritable(String languageType, String singName, Long playTimes) {

        this.set(languageType, singName, playTimes);
    };

    public void set(String languageType, String singName, Long playTimes) {

        this.languageType = languageType;
        this.singName = singName;
        this.playTimes = playTimes;
    }

    public String getLanguageType() {
        return languageType;
    }

    public String getSingName() {
        return singName;
    }

    public Long getPlayTimes() {
        return playTimes;
    }

    @Override
    public void readFields(DataInput in) throws IOException {
        this.languageType = in.readUTF();
        this.singName = in.readUTF();
        this.playTimes = in.readLong();
    }

    @Override
    public void write(DataOutput out) throws IOException {
        out.writeUTF(languageType);
        out.writeUTF(singName);
        out.writeLong(playTimes);

    }

    @Override
    public int compareTo(TopKeyWritable o) {

        // 加个负号倒排序
        return -(this.getPlayTimes().compareTo(o.getPlayTimes()));
    }

    @Override
    public String toString() {
        return languageType + "\t" + singName + "\t" + playTimes;
    }

    @Override
    public int hashCode() {
        final int prime = 31;
        int result = 1;
        result = prime * result
                + ((languageType == null) ? 0 : languageType.hashCode());
        result = prime * result
                + ((playTimes == null) ? 0 : playTimes.hashCode());
        result = prime * result
                + ((singName == null) ? 0 : singName.hashCode());
        return result;
    }

    @Override
    public boolean equals(Object obj) {
        if (this == obj)
            return true;
        if (obj == null)
            return false;
        if (getClass() != obj.getClass())
            return false;
        TopKeyWritable other = (TopKeyWritable) obj;
        if (languageType == null) {
            if (other.languageType != null)
                return false;
        } else if (!languageType.equals(other.languageType))
            return false;
        if (playTimes == null) {
            if (other.playTimes != null)
                return false;
        } else if (!playTimes.equals(other.playTimes))
            return false;
        if (singName == null) {
            if (other.singName != null)
                return false;
        } else if (!singName.equals(other.singName))
            return false;
        return true;
    }

}



你可能感兴趣的:(topkey)