Hadoop壓縮技術的概念

本篇內容主要講解“Hadoop壓縮技術的概念”，感興趣的朋友不妨來看看。本文介紹的方法操作簡單快捷，實用性強。下面就讓小編來帶大家學習“Hadoop壓縮技術的概念”吧!

1 概述

壓縮策略和原則

2 MR 支持的壓縮編碼

為了支持多種壓縮/解壓縮算法，Hadoop 引入了編碼/解碼器，如下表所示。

壓縮性能的比較

3 壓縮方式選擇

3.1 Gzip 壓縮

3.2 Bzip2 壓縮

3.3 Lzo 壓縮

3.4 Snappy 壓縮

4 壓縮位置選擇

5 壓縮參數配置

6 壓縮實操案例

6.1 數據流的壓縮和解壓縮

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 輸出端采用壓縮

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);
        // 設置map端輸出壓縮方式
        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 輸出端采用壓縮

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]));
        // 設置reduce端輸出壓縮開啟
        FileOutputFormat.setCompressOutput(job, true);
        // 設置壓縮的方式
        FileOutputFormat.setOutputCompressorClass(job, BZip2Codec.class);
        boolean result = job.waitForCompletion(true);
        System.exit(result ? 0 : 1);
    }
}

到此，相信大家對“Hadoop壓縮技術的概念”有了更深的了解，不妨來實際操作一番吧！這里是億速云網站，更多相關內容可以進入相關頻道進行查詢，關注我們，繼續學習！