读取hadoop Sequence格式的文件的代码

http://blog.csdn.net/longerandlonger/article/details/8482477

[java] view plaincopy

1. public static void main(String[] args) {  
2.     org.apache.hadoop.io.SequenceFile.Reader reader = null;  
3.     java.io.FileOutputStream fos = null;  
4.     try {  
5.         String uri = "file:///D:/attempt_201212181734_2923950_r_000000_0";  
6.         org.apache.hadoop.conf.Configuration conf = new org.apache.hadoop.conf.Configuration();  
7.         FileSystem fs = FileSystem.get(conf);  
8.         Path path = new Path(uri);  
9.         reader = new org.apache.hadoop.io.SequenceFile.Reader(fs, path, conf);  
10.         Writable key = (Writable) org.apache.hadoop.util.ReflectionUtils.newInstance(reader.getKeyClass(), conf);  
11.         Writable value = (Writable) org.apache.hadoop.util.ReflectionUtils.newInstance(reader.getValueClass(), conf);  
12.           
13.         int n=0;  
14.         while(reader.next(key, value)){  
15.             /* 如果解析出是乱码，尝试用户UTF8转码 */  
16.             //String valueStr = new String(value.toString().getBytes("ISO8859_1"),"GB2312");  
17.             System.out.println(value.toString());  
18.         }  
19.     } catch (Exception e) {  
20.         e.printStackTrace();  
21.     } finally {  
22.         IOUtils.closeStream(reader);  
23.         IOUtils.closeStream(fos);  
24.     }  
25. }