Hadoop压缩技术的概念

本篇内容主要讲解“Hadoop压缩技术的概念”，感兴趣的朋友不妨来看看。本文介绍的方法操作简单快捷，实用性强。下面就让小编来带大家学习“Hadoop压缩技术的概念”吧!

1 概述

压缩策略和原则

2 MR 支持的压缩编码

为了支持多种压缩/解压缩算法，Hadoop 引入了编码/解码器，如下表所示。

压缩性能的比较

3 压缩方式选择

3.1 Gzip 压缩

3.2 Bzip2 压缩

3.3 Lzo 压缩

3.4 Snappy 压缩

4 压缩位置选择

5 压缩参数配置

6 压缩实操案例

6.1 数据流的压缩和解压缩

package com.djm.mapreduce.zip;

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.IOUtils;
import org.apache.hadoop.io.compress.CompressionCodec;
import org.apache.hadoop.io.compress.CompressionCodecFactory;
import org.apache.hadoop.io.compress.CompressionInputStream;
import org.apache.hadoop.io.compress.CompressionOutputStream;
import org.apache.hadoop.util.ReflectionUtils;

import java.io.*;

public class CompressUtils {
    public static void main(String[] args) throws IOException, ClassNotFoundException {
        compress(args[0], args[1]);
        decompress(args[0]);
    }

    private static void decompress(String path) throws IOException {
        CompressionCodecFactory factory = new CompressionCodecFactory(new Configuration());
        CompressionCodec codec = (CompressionCodec) factory.getCodec(new Path(path));
        if (codec == null) {
            System.out.println("cannot find codec for file " + path);
            return;
        }
        CompressionInputStream cis = codec.createInputStream(new FileInputStream(new File(path)));
        FileOutputStream fos = new FileOutputStream(new File(path + ".decoded"));
        IOUtils.copyBytes(cis, fos, 1024);
        cis.close();
        fos.close();
    }

    private static void compress(String path, String method) throws IOException, ClassNotFoundException {
        FileInputStream fis = new FileInputStream(new File(path));
        Class codecClass  = Class.forName(method);
        CompressionCodec codec  = (CompressionCodec) ReflectionUtils.newInstance(codecClass, new Configuration());
        FileOutputStream fos = new FileOutputStream(new File(path + codec.getDefaultExtension()));
        CompressionOutputStream cos = codec.createOutputStream(fos);
        IOUtils.copyBytes(fis, cos, 1024);
        cos.close();
        fos.close();
        fis.close();
    }
}

6.2 Map 输出端采用压缩

package com.djm.mapreduce.wordcount;

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.IntWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.io.compress.BZip2Codec;
import org.apache.hadoop.io.compress.CompressionCodec;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;

import java.io.IOException;

public class WcDriver {
    public static void main(String[] args) throws IOException, ClassNotFoundException, InterruptedException {
        Configuration configuration = new Configuration();
        configuration.setBoolean("mapreduce.map.output.compress", true);
        // 设置map端输出压缩方式
        configuration.setClass("mapreduce.map.output.compress.codec", BZip2Codec.class, CompressionCodec.class);
        Job job = Job.getInstance(configuration);
        job.setJarByClass(WcDriver.class);
        job.setMapperClass(WcMapper.class);
        job.setReducerClass(WcReduce.class);
        job.setMapOutputKeyClass(Text.class);
        job.setMapOutputValueClass(IntWritable.class);
        job.setOutputKeyClass(Text.class);
        job.setOutputValueClass(IntWritable.class);
        FileInputFormat.setInputPaths(job, new Path(args[0]));
        FileOutputFormat.setOutputPath(job, new Path(args[1]));
        boolean result = job.waitForCompletion(true);
        System.exit(result ? 0 : 1);
    }
}

6.3 Reduce 输出端采用压缩

package com.djm.mapreduce.wordcount;

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.IntWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.io.compress.BZip2Codec;
import org.apache.hadoop.io.compress.CompressionCodec;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;

import java.io.IOException;

public class WcDriver {
    public static void main(String[] args) throws IOException, ClassNotFoundException, InterruptedException {
        Configuration configuration = new Configuration();
        Job job = Job.getInstance(configuration);
        job.setJarByClass(WcDriver.class);
        job.setMapperClass(WcMapper.class);
        job.setReducerClass(WcReduce.class);
        job.setMapOutputKeyClass(Text.class);
        job.setMapOutputValueClass(IntWritable.class);
        job.setOutputKeyClass(Text.class);
        job.setOutputValueClass(IntWritable.class);
        FileInputFormat.setInputPaths(job, new Path(args[0]));
        FileOutputFormat.setOutputPath(job, new Path(args[1]));
        // 设置reduce端输出压缩开启
        FileOutputFormat.setCompressOutput(job, true);
        // 设置压缩的方式
        FileOutputFormat.setOutputCompressorClass(job, BZip2Codec.class);
        boolean result = job.waitForCompletion(true);
        System.exit(result ? 0 : 1);
    }
}

到此，相信大家对“Hadoop压缩技术的概念”有了更深的了解，不妨来实际操作一番吧！这里是亿速云网站，更多相关内容可以进入相关频道进行查询，关注我们，继续学习！