Hadoop自定义输出的代码 第2页

bsp;              
                //private Log LOG = LogFactory.getLog(ZRecordReader.class);//日志写入系统，可加可不加   
                               
                @Override   
                public void initialize(InputSplit split, TaskAttemptContext context)   
                        throws IOException, InterruptedException {   
                    // 初始化函数   
                       
                    FileSplit inputsplit = (FileSplit)split;   
                    start = inputsplit.getStart();                      //得到此分片开始位置   
                    end   = start + inputsplit.getLength();//结束此分片位置   
                    final Path file = inputsplit.getPath();   
               
                    // 打开文件   
                    FileSystem fs = file.getFileSystem(context.getConfiguration());   
                    FSDataInputStream fileIn = fs.open(inputsplit.getPath());   
                       
                     
                    //将文件指针移动到当前分片，因为每次默认打开文件时，其指针指向开头   
                    fileIn.seek(start);   
                       
                    in = new LineReader(fileIn, context.getConfiguration());   
       
                    if (start != 0)    
                    {   
                      System.out.println("4");    
                       //如果这不是第一个分片，那么假设第一个分片是0——4，那么，第4个位置已经被读取，则需要跳过4，否则会产生读入错误，因为你回头又去读之前读过的地方   
                   start += in.readLine(new Text(), 0, maxBytesToConsume(start));   
                    }   
                    pos = start;   
          

上一页  [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] 下一页