Hadoop 减少输入记录=0答案

【问题标题】：Hadoop Reduce input records=0Hadoop 减少输入记录=0
【发布时间】：2017-01-28 18:00:43
【问题描述】：

我是 Hadoop 新手，我的 map-reduce 代码可以工作，但它不会产生任何输出。这是map-reduce的信息：

16/09/20 13:11:40 INFO mapred.JobClient: Job complete: job_201609081210_0078
16/09/20 13:11:40 INFO mapred.JobClient: Counters: 28
16/09/20 13:11:40 INFO mapred.JobClient:   Map-Reduce Framework
16/09/20 13:11:40 INFO mapred.JobClient:     Spilled Records=0
16/09/20 13:11:40 INFO mapred.JobClient:     Map output materialized bytes=1362
16/09/20 13:11:40 INFO mapred.JobClient:     Reduce input records=0
16/09/20 13:11:40 INFO mapred.JobClient:     Virtual memory (bytes)   snapshot=466248720384
16/09/20 13:11:40 INFO mapred.JobClient:     Map input records=852032443
16/09/20 13:11:40 INFO mapred.JobClient:     SPLIT_RAW_BYTES=29964
16/09/20 13:11:40 INFO mapred.JobClient:     Map output bytes=0
16/09/20 13:11:40 INFO mapred.JobClient:     Reduce shuffle bytes=1362
16/09/20 13:11:40 INFO mapred.JobClient:     Physical memory (bytes) snapshot=57472311296
16/09/20 13:11:40 INFO mapred.JobClient:     Reduce input groups=0
16/09/20 13:11:40 INFO mapred.JobClient:     Combine output records=0
16/09/20 13:11:40 INFO mapred.JobClient:     Reduce output records=0
16/09/20 13:11:40 INFO mapred.JobClient:     Map output records=0
16/09/20 13:11:40 INFO mapred.JobClient:     Combine input records=0
16/09/20 13:11:40 INFO mapred.JobClient:     CPU time spent (ms)=2375210
16/09/20 13:11:40 INFO mapred.JobClient:     Total committed heap usage (bytes)=47554494464
16/09/20 13:11:40 INFO mapred.JobClient:   File Input Format Counters
16/09/20 13:11:40 INFO mapred.JobClient:     Bytes Read=15163097088
16/09/20 13:11:40 INFO mapred.JobClient:   FileSystemCounters
16/09/20 13:11:40 INFO mapred.JobClient:     HDFS_BYTES_READ=15163127052
16/09/20 13:11:40 INFO mapred.JobClient:     FILE_BYTES_WRITTEN=13170190
16/09/20 13:11:40 INFO mapred.JobClient:     FILE_BYTES_READ=6
16/09/20 13:11:40 INFO mapred.JobClient:   Job Counters
16/09/20 13:11:40 INFO mapred.JobClient:     Launched map tasks=227
16/09/20 13:11:40 INFO mapred.JobClient:     Launched reduce tasks=1
16/09/20 13:11:40 INFO mapred.JobClient:     SLOTS_MILLIS_REDUCES=759045
16/09/20 13:11:40 INFO mapred.JobClient:     Total time spent by all reduces waiting after reserving slots (ms)=0
16/09/20 13:11:40 INFO mapred.JobClient:     SLOTS_MILLIS_MAPS=1613259
16/09/20 13:11:40 INFO mapred.JobClient:     Total time spent by all maps waiting after reserving slots (ms)=0
16/09/20 13:11:40 INFO mapred.JobClient:     Data-local map tasks=227
16/09/20 13:11:40 INFO mapred.JobClient:   File Output Format Counters
16/09/20 13:11:40 INFO mapred.JobClient:     Bytes Written=0

这是启动 mapreduce 作业的代码：

import java.io.File;
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.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;

public class mp{

public static void main(String[] args) throws Exception {

    Job job1 = new Job();
    job1.setJarByClass(mp.class);
    FileInputFormat.addInputPath(job1, new Path(args[0]));                  
    String oFolder = args[0] + "/output";
    FileOutputFormat.setOutputPath(job1, new Path(oFolder));
    job1.setMapperClass(TransMapper1.class);
    job1.setReducerClass(TransReducer1.class);
    job1.setMapOutputKeyClass(LongWritable.class);
    job1.setMapOutputValueClass(DnaWritable.class);
    job1.setOutputKeyClass(LongWritable.class);
    job1.setOutputValueClass(Text.class);
}
}

这里是映射器类（TransMapper1）：

import java.io.IOException;
import java.util.StringTokenizer;
import org.apache.hadoop.io.DoubleWritable;
import org.apache.hadoop.io.IntWritable;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Mapper;

public class TransMapper1 extends  Mapper<LongWritable, Text, LongWritable, DnaWritable> {

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

        String line = value.toString();
        StringTokenizer tokenizer = new StringTokenizer(line);
        LongWritable bamWindow = new LongWritable(Long.parseLong(tokenizer.nextToken()));
        LongWritable read = new LongWritable(Long.parseLong(tokenizer.nextToken()));
        LongWritable refWindow = new LongWritable(Long.parseLong(tokenizer.nextToken()));
        IntWritable chr = new IntWritable(Integer.parseInt(tokenizer.nextToken()));
        DoubleWritable dist = new DoubleWritable(Double.parseDouble(tokenizer.nextToken()));
        DnaWritable dnaW = new DnaWritable(bamWindow,read,refWindow,chr,dist);
        context.write(bamWindow,dnaW);
    }
}

这是Reducer类（TransReducer1）：

import java.io.IOException;
import java.util.ArrayList;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Reducer;

public class TransReducer1 extends Reducer<LongWritable, DnaWritable, LongWritable, Text> {

@Override
 public void reduce(LongWritable key, Iterable<DnaWritable> values, Context context) throws IOException, InterruptedException {

ArrayList<DnaWritable> list = new ArrayList<DnaWritable>();
double minDist = Double.MAX_VALUE;
    for (DnaWritable value : values) {
            long bamWindow = value.getBamWindow().get();
            long read = value.getRead().get();
            long refWindow = value.getRefWindow().get();
            int chr = value.getChr().get();
            double dist = value.getDist().get();
            if (dist > minDist)
                continue;
            else
            if (dist < minDist)
                 list.clear();
            list.add(new DnaWritable(bamWindow,read,refWindow,chr,dist));
            minDist = Math.min(minDist, value.getDist().get());
        }
        for(int i = 0; i < list.size(); i++){
            context.write(new LongWritable(list.get(i).getRead().get()),new Text(new DnaWritable(list.get(i).getBamWindow(),list.get(i).getRead(),list.get(i).getRefWindow(),list.get(i).getChr(),list.get(i).getDist()).toString()));
        }
    }
}

这是 DnaWritable 类（我没有把导入部分缩短一点）：

public class DnaWritable implements Writable {
    LongWritable bamWindow;
    LongWritable read;
    LongWritable refWindow;
    IntWritable chr;
    DoubleWritable dist;

    public DnaWritable(LongWritable bamWindow, LongWritable read, LongWritable refWindow, IntWritable chr, DoubleWritable dist){

    this.bamWindow = bamWindow;
    this.read = read;
    this.refWindow = refWindow;
    this.chr = chr;
    this.dist = dist;
}

public DnaWritable(long bamWindow, long read, long refWindow, int chr, double dist){
    this.bamWindow = new LongWritable(bamWindow);
    this.read = new LongWritable(read);
    this.refWindow = new LongWritable(refWindow);
    this.chr = new IntWritable(chr);
    this.dist = new DoubleWritable(dist);
}

@Override
public void write(DataOutput dataOutput) throws IOException {
    bamWindow.write(dataOutput);
    read.write(dataOutput);
    refWindow.write(dataOutput);
    chr.write(dataOutput);
    dist.write(dataOutput);
}

@Override
public void readFields(DataInput dataInput) throws IOException {
        bamWindow.readFields(dataInput);
        read.readFields(dataInput);
        refWindow.readFields(dataInput);
        chr.readFields(dataInput);
        dist.readFields(dataInput);
    }
}

任何帮助将不胜感激..谢谢

【问题讨论】：

你能提供一些示例数据吗？
是的，样本输入数据会在这里得到答案，提供两三行输入
谢谢你们...我刚刚在我的一个输入文件中发现了一个缺陷...我希望这能解决我的问题。谢谢

标签： java hadoop mapreduce mapper reducers

【解决方案1】：

我认为您没有在 DnaWritable 类中正确实现 write(DataOutput out) 和 readFields(DataInput in) 方法。

【讨论】：

我也放了 DnaWritable 代码。请看一下。谢谢

【解决方案2】：

您能否将您的 DnaWritable 类更改为并进行测试。（处理 NPE）

package com.hadoop.intellipaat;

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

import org.apache.hadoop.io.Writable;

public class DnaWritable implements Writable {

    private Long bamWindow;
    private Long read;
    private Long refWindow;
    private Integer chr;
    private Double dist;

    public DnaWritable(Long bamWindow, Long read, Long refWindow, Integer chr, Double dist) {
        super();
        this.bamWindow = bamWindow;
        this.read = read;
        this.refWindow = refWindow;
        this.chr = chr;
        this.dist = dist;
    }

    @Override
    public void write(DataOutput out) throws IOException {
        out.writeLong(bamWindow);
        out.writeLong(read);
        out.writeLong(refWindow);
        out.writeInt(chr);
        out.writeDouble(dist);
    }

    @Override
    public void readFields(DataInput in) throws IOException {
        this.bamWindow = in.readLong();
        this.read = in.readLong();
        this.refWindow = in.readLong();
        this.chr = in.readInt();
        this.dist = in.readDouble();
    }

}

【讨论】：

最好避免盒装图元。我建议将字段更改为它们的原始类型。我这么说只是因为看起来空值不是 OPs 字段的可能值。
我改变了实现，但到目前为止没有运气
@Hamid_UMB 请提供一些示例数据？？这样我就可以检查了。

【解决方案3】：

考虑也实现 ComparableWritable 如下，同时添加无参数构造函数。

public class DnaWritable implements Writable WritableComparable<DnaWritable>  {

 //Consider add a non-args constructor
 public DnaWritable(){
 }

    //Add this static method as well
 public static DnaWritable read(DataInput in) throws IOException {
        DnaWritable dnaWritable = new DnaWritable();
        dnaWritable.readFields(in);
        return dnaWritable;
 }

 @Override
 public int compareTo(DnaWritable dnaWritable) {
      //Put your comparison logic there.
 }

}

如果仍然失败，请查看 log4.properties，以便您查看是否有任何您没有看到的潜在错误。

src/main/resources

hadoop.root.logger=DEBUG, console
log4j.rootLogger=INFO, stdout

# Direct log messages to stdout
log4j.appender.stdout=org.apache.log4j.ConsoleAppender
log4j.appender.stdout.Target=System.out
log4j.appender.stdout.layout=org.apache.log4j.PatternLayout
log4j.appender.stdout.layout.ConversionPattern=%d{yyyy-MM-dd HH:mm:ss} %-5p %c{1}:%L - %m%n

【讨论】：

当我试图实现 WritableComparable 接口时，我遇到了这个错误：错误：不能从静态上下文引用非静态方法 readFields(DataInput) DnaWritable.readFields(in); .. .我该如何解决？？
它应该是 dnaWritable ，小写我不好，应该是最近创建的实例。

【解决方案4】：

我认为您根本没有向集群提交工作。您的主类中没有 job1.submit() 或 job1.waitForCompletion(true)。

////submit the job to hadoop 
if (!job1.waitForCompletion(true))
return;

您的主要方法中也需要进行更正。

Job job1 = new Job();  //new Job() constructor is deprecated now.

下面是创建作业对象的正确方法

Configuration conf = new Configuration();
Job job1 = Job.getInstance(conf, "Your Program name");

【讨论】：