刪除HDFS集羣中的全部空文件和空目錄

一、鏈接HDFS

package api;


import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.FileSystem;

public class Utils {

    public static FileSystem HDFS() throws Exception{
        
        Configuration conf = new Configuration();
        conf.set("fs.defaultFS","hdfs://potter2:9000");
        System.setProperty("HADOOP_USER_NAME", "potter");
        FileSystem fs = FileSystem.get(conf);
        return fs;
         
         
    }
}

二、主要代碼

package api;

import org.apache.hadoop.fs.FileStatus;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.LocatedFileStatus;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.fs.RemoteIterator;
import org.junit.Test;

/**
 * 刪除HDFS集羣中的全部空文件和空目錄
 * @author Administrator
 *
 */
public class Empty {

        @Test
        public void tt() throws Exception {
            Path path = new Path("/");
            Empty1(path);
        }
        
        public static void Empty1(Path path) throws Exception {
            FileSystem fs = Utils.HDFS();
            
            //當是空文件時，判斷當前路徑下有幾個空文件夾
            FileStatus[] listStatus = fs.listStatus(path);
            System.out.println(listStatus.length+"********");
            //當根目錄沒有文件的時候會進入if裏面
            if (listStatus.length == 0) {
                fs.delete(path,true);
            }
            System.out.println("刪除成功xxxxx");
            //迭代器用於遍歷
            RemoteIterator<LocatedFileStatus> listLocatedStatus = fs.listLocatedStatus(path);
        
            while (listLocatedStatus.hasNext()) {

                LocatedFileStatus next = listLocatedStatus.next();
                //輸出文件夾的目錄
                Path currentPath = next.getPath();
                System.out.println(currentPath+"1111111");
                //輸出上面文件夾的父親目錄
                Path parent = next.getPath().getParent();
                System.out.println(parent+"2222222");
                if (next.isDirectory()) {
                    //若是是空文件夾
                    if (fs.listStatus(currentPath).length == 0) {
                        //刪除掉
                        fs.delete(currentPath,true);
                    }else {
                        //不是空文件夾，那麼繼續遍歷
                        if (fs.exists(currentPath)) {
                            Empty1(next.getPath());
                        }
                    }
                //若是是文件
                }else {
                    //獲取文件的長度
                    long fileLength = next.getLen();
                    //當文件是空文件時，刪除
                    if (fileLength ==0) {
                        fs.delete(currentPath,true);
                    }
                }
                int length = fs.listStatus(parent).length;
                if (length ==0) {
                    fs.delete(parent,true);
                }
                
            }
        }
}