电影评分案例之高效TopN

例如:我们要求每部电影的最高评分的前n条记录,按照之前的做法在map端是以电影名为key,MovieBean为value,输出到reduce端,然后分组,将每组数组放入到List集合中按分数高低进行排序,取前n条.
此时我么可以考虑在map端时将MovieBean作为key,输出到缓存区中,让缓存区自动按电影名分区并排序,然后分组,在reduce端我们只需要取出前n条记录即可.这样我们可以避免放入List集合中再排一遍序,大大的减少了运算量.
那么当我们以MovieBean为key是,要想系统识别到是以MovieBean中的电影名分区,排序,分组,我们就需要重写这三个方法,并且还需要重写hadoop的序列化和反序列化方法

代码实现:

import org.apache.hadoop.io.WritableComparable;
import java.io.DataInput;
import java.io.DataOutput;
import java.io.IOException;

/**
 *以MovieBeanTopN为key输出到缓存区中,系统排序时,需要知道排序规则,及序列化和反序列化规则
 * 在MovieBeanTopN中我们可以实现WritableComparable接口,注意加上范型为MovieBeanTopN
 * 然后重写它里面的三个方法:
 * 序列化
 * 反序列化
 * 自定义排序
 */
public class MovieBeanTopN implements WritableComparable<MovieBeanTopN> {

    private String movie;
    private double rate;
    private String timeStamp;
    private String uid;

    /**
     * 自定义toString方法,方便生成的文件便于后续读取切分,还可以减少文件的大小
     * @return
     */
    @Override
    public String toString() {
        return  movie + ", " + rate + ", " + timeStamp + ", " + uid ;
    }
    public String getMovie() {
        return movie;
    }
    public void setMovie(String movie) {
        this.movie = movie;
    }
    public double getRate() {
        return rate;
    }
    public void setRate(double rate) {
        this.rate = rate;
    }
   public String getTimeStamp() {
        return timeStamp;
    }
    public void setTimeStamp(String timeStamp) {
        this.timeStamp = timeStamp;
    }

   public String getUid() {
        return uid;
    }
    public void setUid(String uid) {
        this.uid = uid;
    }


    /**
     * 重写序列化方法
     * @param dataOutput
     * @throws IOException
     */
    @Override
    public void write(DataOutput dataOutput) throws IOException {
        dataOutput.writeUTF(movie);
        dataOutput.writeDouble(rate);
        dataOutput.writeUTF(timeStamp);
        dataOutput.writeUTF(uid);
    }

    /**
     * 重写反序列化方法
     * @param dataInput
     * @throws IOException
     */
    @Override
    public void readFields(DataInput dataInput) throws IOException {
        this.movie = dataInput.readUTF();
        this.rate = dataInput.readDouble();
        this.timeStamp = dataInput.readUTF();
        this.uid = dataInput.readUTF();
    }

    /**
     * 自定义排序方法
     * @param o
     * @return
     */
    @Override
    public int compareTo(MovieBeanTopN o) {
    //先比较电影名是否相同,如果相同就按评分降序,如果电影名不同就按电影名降序
        return o.getMovie().compareTo(this.movie)==0?Double.compare(o.getRate(),this.rate):o.getMovie().compareTo(this.movie);
    }
}
import org.apache.hadoop.io.NullWritable;
import org.apache.hadoop.mapreduce.Partitioner;

/**
 *extends Partitioner类,自定义分区方法,
 * 注意范型类型为map端输出的两个类型
 */
public class MyPartition extends Partitioner<MovieBeanTopN, NullWritable> {
    @Override //以电影名.hashCode() % reduce任务个数
    public int getPartition(MovieBeanTopN key, NullWritable nullWritable, int i) {
        return (key.getMovie().hashCode() & 2147483647) % i;
    }
}
import org.apache.hadoop.io.WritableComparable;
import org.apache.hadoop.io.WritableComparator;

/**
 *extends WritableComparato自定义分组规则
 */
public class MyGropingComparapor extends WritableComparator {
    /**
     * 因为继承的关系,子类会默认调用父类的空参构造,但WritableComparator的空参构造中的参数为     null,所以调用父类对象时需要传入一个对象,避免空指针
     */
    public MyGropingComparapor(){
        super(MovieBeanTopN.class,true);
    }

    /**
     * 自定义分组规则,比较相邻的两对象的movie值是否相同
     * @param a
     * @param b
     * @return
     */
    public int compare(WritableComparable a, WritableComparable b) {
        MovieBeanTopN a1 = (MovieBeanTopN) a;
        MovieBeanTopN b1 = (MovieBeanTopN) b;
        return a1.getMovie().compareTo(b1.getMovie());
    }
}
import com.alibaba.fastjson.JSON;
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.output.FileOutputFormat;
import java.io.IOException;

/**
 *以MovieBeanTopN为key,放入缓存区中让他自动分区,排序并分组,这样可以高效的处理数据,
 * 但需要我们自定义分区规则,分组规则,排序规则,重写序列化和反序列化方法(上面已经定义)
 */
public class MovieTopN{

    static class MovieTopNMap extends Mapper<LongWritable, Text,MovieBeanTopN, NullWritable> {

        /**
         * 以MovieBeanTopN为key,NullWritable为value输出,在缓存区就会自动按我们重写的规则分区并排序
         * @param key
         * @param value
         * @param context
         * @throws IOException
         * @throws InterruptedException
         */
    @Override
    protected void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException {
        try {
            String s = value.toString();
            MovieBeanTopN mb = JSON.parseObject(s, MovieBeanTopN.class);

            context.write(mb,NullWritable.get());
        } catch (Exception e) {
            e.printStackTrace();
        }
    }
}
    static class MovieTopNReduce extends Reducer<MovieBeanTopN, NullWritable,MovieBeanTopN, NullWritable>{
        /**
         *获取排序后的key(MovieBeanTopN)和value(<NullWritable>),遍历value的迭代器,因为索引的关系,会将key和value联系起来
         * 遍历value就能得到它对应的key,所以可以在遍历value时直接输出分好组的key和NullWritable
         * @param key
         * @param values
         * @param context
         * @throws IOException
         * @throws InterruptedException
         */
        @Override
        protected void reduce(MovieBeanTopN key, Iterable<NullWritable> values, Context context) throws IOException, InterruptedException {

            int count = 0;
            for (NullWritable value : values) {
                count++;
                context.write(key,NullWritable.get());
                if (count==3){
                    return;
                }
            }
        }
    }

    public static void main(String[] args) throws Exception {
        //配置对象
        Configuration conf = new Configuration();
        //配置任务对象
        Job job = Job.getInstance(conf, "cont_top1");
        //导入任务类
        job.setMapperClass(MovieTopNMap.class);
        job.setReducerClass(MovieTopNReduce.class);
        //设置输出类型
        job.setOutputKeyClass(MovieBeanTopN.class);
        job.setOutputValueClass(NullWritable.class);
        //导入我们自定义分区的类
        job.setPartitionerClass(MyPartition.class);
        //导入我们自定义分组的类
        job.setGroupingComparatorClass(MyGropingComparapor.class);
        //设置reduce的任务个数
        job.setNumReduceTasks(2);

        //设置输入输出路径
        FileInputFormat.setInputPaths(job, new Path("D:\\txt\\mrdata\\movie\\input"));
        FileOutputFormat.setOutputPath(job, new Path("D:\\txt\\mrdata\\movie\\output8"));
        //  提交任务  等待程序执行完毕   返回值是否成功
        boolean b = job.waitForCompletion(true);
    }
}