文件的压缩有两大好处：1、可以减少存储文件所需要的磁盘空间；2、可以加速数据在网络和磁盘上的传输。尤其是在处理大数据时，这两大好处是相当重要的。
下面是一个使用gzip工具压缩文件的例子。将文件/user/hadoop/aa.txt进行压缩，压缩后为/user/hadoop/text.gz
1 package com.hdfs;
2
3 import java.io.IOException;
4 import java.io.InputStream;
5 import java.io.OutputStream;
6 import java.net.URI;
7
8 import org.apache.hadoop.conf.Configuration;
9 import org.apache.hadoop.fs.FSDataInputStream;
10 import org.apache.hadoop.fs.FSDataOutputStream;
11 import org.apache.hadoop.fs.FileSystem;
12 import org.apache.hadoop.fs.Path;
13 import org.apache.hadoop.io.IOUtils;
14 import org.apache.hadoop.io.compress.CompressionCodec;
15 import org.apache.hadoop.io.compress.CompressionCodecFactory;
16 import org.apache.hadoop.io.compress.CompressionInputStream;
17 import org.apache.hadoop.io.compress.CompressionOutputStream;
18 import org.apache.hadoop.util.ReflectionUtils;
19
20 public class CodecTest {
21 //压缩文件
22 public static void compress(String codecClassName) throws Exception{
23 Class<?> codecClass = Class.forName(codecClassName);
24 Configuration conf = new Configuration();
25 FileSystem fs = FileSystem.get(conf);
26 CompressionCodec codec = (CompressionCodec)ReflectionUtils.newInstance(codecClass, conf);
27 //指定压缩文件路径
28 FSDataOutputStream outputStream = fs.create(new Path(“/user/hadoop/text.gz”));
29 //指定要被压缩的文件路径
30 FSDataInputStream in = fs.open(new Path(“/user/hadoop/aa.txt”));
31 //创建压缩输出流
32 CompressionOutputStream out = codec.createOutputStream(outputStream);
33 IOUtils.copyBytes(in, out, conf);
34 IOUtils.closeStream(in);
35 IOUtils.closeStream(out);
36 }
37
38 //解压缩
39 public static void uncompress(String fileName) throws Exception{
40 Class<?> codecClass = Class.forName(“org.apache.hadoop.io.compress.GzipCodec”);
41 Configuration conf = new Configuration();
42 FileSystem fs = FileSystem.get(conf);
43 CompressionCodec codec = (CompressionCodec)ReflectionUtils.newInstance(codecClass, conf);
44 FSDataInputStream inputStream = fs.open(new Path(“/user/hadoop/text.gz”));
45 //把text文件里到数据解压，然后输出到控制台
46 InputStream in = codec.createInputStream(inputStream);
47 IOUtils.copyBytes(in, System.out, conf);
48 IOUtils.closeStream(in);
49 }
50
51 //使用文件扩展名来推断二来的codec来对文件进行解压缩
52 public static void uncompress1(String uri) throws IOException{
53 Configuration conf = new Configuration();
54 FileSystem fs = FileSystem.get(URI.create(uri), conf);
55
56 Path inputPath = new Path(uri);
57 CompressionCodecFactory factory = new CompressionCodecFactory(conf);
58 CompressionCodec codec = factory.getCodec(inputPath);
59 if(codec == null){
60 System.out.println(“no codec found for “ + uri);
61 System.exit(1);
62 }
63 String outputUri = CompressionCodecFactory.removeSuffix(uri, codec.getDefaultExtension());
64 InputStream in = null;
65 OutputStream out = null;
66 try {
67 in = codec.createInputStream(fs.open(inputPath));
68 out = fs.create(new Path(outputUri));
69 IOUtils.copyBytes(in, out, conf);
70 } finally{
71 IOUtils.closeStream(out);
72 IOUtils.closeStream(in);
73 }
74 }
75
76 public static void main(String[] args) throws Exception {
77 //compress(“org.apache.hadoop.io.compress.GzipCodec”);
78 //uncompress(“text”);
79 uncompress1(“hdfs://master:9000/user/hadoop/text.gz”);
80 }
81
82 }首先执行77行进行压缩，压缩后执行第78行进行解压缩，这里解压到标准输出，所以执行78行会再控制台看到文件/user/hadoop /aa.txt的内容。如果执行79行的话会将文件解压到/user/hadoop/text,他是根据/user/hadoop/text.gz的扩展 名判断使用哪个解压工具进行解压的。解压后的路径就是去掉扩展名。
进行文件压缩后，在执行命令./hadoop fs -ls /user/hadoop/查看文件信息，如下：
1 [hadoop@master bin]$ ./hadoop fs -ls /user/hadoop/
2 Found 7 items
3 -rw-r–r– 3 hadoop supergroup 76805248 2013-06-17 23:55 /user/hadoop/aa.mp4
4 -rw-r–r– 3 hadoop supergroup 520 2013-06-17 22:29 /user/hadoop/aa.txt
5 drwxr-xr-x - hadoop supergroup 0 2013-06-16 17:19 /user/hadoop/input
6 drwxr-xr-x - hadoop supergroup 0 2013-06-16 19:32 /user/hadoop/output
7 drwxr-xr-x - hadoop supergroup 0 2013-06-18 17:08 /user/hadoop/test
8 drwxr-xr-x - hadoop supergroup 0 2013-06-18 19:45 /user/hadoop/test1
9 -rw-r–r– 3 hadoop supergroup 46 2013-06-19 20:09 /user/hadoop/text.gz第4行为压缩之前的文件，大小为520个字节。第9行为压缩后的文件，大小为46个字节。由此可以看出上面讲的压缩的两大好处了。
