HDFS 客户端操作

1.HDFS客户端环境准备
(1)根据自己的电脑的操作系统拷贝对应的编译后的hadoop jar包到非中文路径(例如我的MAC:/Users/luomingkui/hadoop/hadoop-2.7.2)。
(2)配置MAC环境变量
(3)创建一个Maven工程HdfsClientDemo
(4)导入相应的依赖
<dependencies>
  <dependency>
         <groupId>junit</groupId>
         <artifactId>junit</artifactId>
         <version>RELEASE</version>
  </dependency>
  <dependency>
         <groupId>org.apache.logging.log4j</groupId>
         <artifactId>log4j-core</artifactId>
         <version>2.8.2</version>
  </dependency>
  <dependency>
         <groupId>org.apache.hadoop</groupId>
         <artifactId>hadoop-common</artifactId>
         <version>2.7.2</version>
  </dependency>
  <dependency>
         <groupId>org.apache.hadoop</groupId>
         <artifactId>hadoop-client</artifactId>
         <version>2.7.2</version>
  </dependency>
  <dependency>
         <groupId>org.apache.hadoop</groupId>
         <artifactId>hadoop-hdfs</artifactId>
         <version>2.7.2</version>
  </dependency>
</dependencies>
(5)创建包名:com.luomk.hdfs
(6)创建HdfsClient类
public class HdfsClient {
    @Test
    public void testMkdirs() throws IOException, InterruptedException, URISyntaxException {
        // 1 获取文件系统
        Configuration configuration = new Configuration();
        // 配置在集群上运行
        // configuration.set("fs.defaultFS", "hdfs://hadoop103:9000");
        // FileSystem fs = FileSystem.get(configuration);
        FileSystem fs = FileSystem.get(new URI("hdfs://hadoop103:9000"), configuration, "luomk");
        // 2 创建目录
        fs.mkdirs(new Path("/0115/daxian/luomk"));
        // 3 关闭资源
        fs.close();
    }
}
 (7)执行程序
    运行时需要配置用户名称
HDFS 客户端操作
    客户端去操作HDFS时,是有一个用户身份的默认情况下,HDFS客户端API会从JVM中获取一个参数来作为自己的用户身份:-DHADOOP_USER_NAME = luomk,luomk为用户名称。
(8)注意:如果eclipse打印不出日志,在控制台上只显示
    1.log4j:WARN记录器(org.apache.hadoop.util.Shell)找不到appender。  
    2.log4j:WARN请正确初始化log4j系统。  
    3.log4j:WARN请参阅http://logging.apache.org/log4j/1.2/faq.html#noconfig了解更多信息。
需要在项目的的src / main /资源目录下,新建一个文件,命名为“log4j.properties”,在文件中填入
log4j.rootLogger =信息,标准输出
log4j.appender.stdout = org.apache.log4j.ConsoleAppender
log4j.appender.stdout.layout = org.apache.log4j.PatternLayout
log4j.appender.stdout.layout.ConversionPattern =%d%p [%c] - %m%n
log4j.appender.logfile = org.apache.log4j.FileAppender
log4j.appender.logfile.File =目标/ spring.log
log4j.appender.logfile.layout = org.apache.log4j.PatternLayout
log4j.appender.logfile.layout.ConversionPattern =%d%p [%c] - %m%n
 
2.HDFS的API操作
(1)HDFS文件上传(测试参数优先级)
    (1)编写源代码
@Test
public void testPut() throws Exception {
    //获取配置文件
    Configuration configuration = new Configuration();
    configuration.set("dfs.replication", "2");
    //获取文件系统
    FileSystem fs = FileSystem.get(new URI("hdfs://hadoop103:9000"), configuration, "luomk");
    //上传文件
    fs.copyFromLocalFile(new Path("/Users/luomingkui/Downloads/wc.input"), new Path("/luomk/wc.input"));
    System.out.println("Over!");
    //关闭资源
    fs.close();
}
    (2)将hdfs-site.xml拷贝到项目的根目录下
<?xml version="1.0" encoding="UTF-8"?>
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?>
<configuration>
    <property>
        <name>dfs.replication</name>
        <value>1</value>
    </property>
</configuration>
    (3)参数优先级
    参数优先级排序:(1)客户端代码中设置的值>(2)classpath下的用户自定义配置文件>(3)然后是服务器的默认配置

(2)HDFS文件下载
  @Test
    public void testCopyToLocalFile() throws IOException, InterruptedException, URISyntaxException{
        // 1 获取文件系统
        Configuration configuration = new Configuration();
        FileSystem fs = FileSystem.get(new URI("hdfs://hadoop103:9000"), configuration, "luomk");
        // 2 执行下载操作
        // boolean delSrc 指是否将原文件删除
        // Path src 指要下载的文件路径
        // Path dst 指将文件下载到的路径
        // boolean useRawLocalFileSystem 是否开启文件校验
       fs.copyToLocalFile(false,new Path("/luomk/wc.input"), new Path("/Users/luomingkui/Downloads/"));
        // 3 关闭资源
        fs.close();
    }
(3)HDFS文件夹删除
@Test
public void testDelete() throws IOException, InterruptedException, URISyntaxException{
    // 1 获取文件系统
    Configuration configuration = new Configuration();
    FileSystem fs = FileSystem.get(new URI("hdfs://hadoop103:9000"), configuration, "luomk");
    // 2 执行删除
    fs.delete(new Path("/over"), true);
    // 3 关闭资源
    fs.close();
}
(4)HDFS文件名更改
@Test
public void testRename() throws IOException, InterruptedException, URISyntaxException{
    // 1 获取文件系统
    Configuration configuration = new Configuration();
    FileSystem fs = FileSystem.get(new URI("hdfs://hadoop103:9000"), configuration, "luomk");
    // 2 修改文件名称
    fs.rename(new Path("/luomk/wc.input"), new Path("/luomk/wc1.input"));
    // 3 关闭资源
    fs.close();
}
 (5) HDFS文件详情查看
查看文件名称、权限、长度、块信息
@Test
    public void testListFiles() throws IOException, InterruptedException, URISyntaxException{
        // 1获取文件系统
        Configuration configuration = new Configuration();
        FileSystem fs = FileSystem.get(new URI("hdfs://hadoop103:9000"), configuration, "luomk");
        
        // 2 获取文件详情
        RemoteIterator<LocatedFileStatus> listFiles = fs.listFiles(new Path("/"), true);
        
        while(listFiles.hasNext()){
            LocatedFileStatus status = listFiles.next();
            
            // 输出详情
            // 文件名称
            System.out.println(status.getPath().getName());
            // 长度
            System.out.println(status.getLen());
            // 权限
            System.out.println(status.getPermission());
            // z组
            System.out.println(status.getGroup());
            
            // 获取存储的块信息
            BlockLocation[] blockLocations = status.getBlockLocations();
            
            for (BlockLocation blockLocation : blockLocations) {
                
                // 获取块存储的主机节点
                String[] hosts = blockLocation.getHosts();
                
                for (String host : hosts) {
                    System.out.println(host);
                }
            }
            System.out.println("----------------分割线-----------");
        }
    }
 (6) HDFS文件和文件夹判断
 @Test
    public void testListStatus() throws IOException, InterruptedException, URISyntaxException{
        // 1 获取文件配置信息
        Configuration configuration = new Configuration();
        FileSystem fs = FileSystem.get(new URI("hdfs://hadoop103:9000"), configuration, "luomk");
        // 2 判断是文件还是文件夹
        FileStatus[] listStatus = fs.listStatus(new Path("/"));
        for (FileStatus fileStatus : listStatus) {
            // 如果是文件
            if (fileStatus.isFile()) {
                System.out.println("f:"+fileStatus.getPath().getName());
            }else {
                System.out.println("d:"+fileStatus.getPath().getName());
            }
        }
        // 3 关闭资源
        fs.close();
    }
 
   
3. HDFS的I/O流操作
(1) HDFS文件上传
public static void put() throws Exception {
    //1.获取配置信息
    Configuration configuration = new Configuration();
    //2.获取文件系统
    FileSystem fs = FileSystem.get(new URI("hdfs://hadoop103:9000"), configuration, "luomk");
    //3.创建输入流
    FileInputStream fis = new FileInputStream("/Users/luomingkui/Downloads/wc.input");
    //4.创建输出流
    FSDataOutputStream fos = fs.create(new Path("/luomk/wc.input"));
    //5.流的对接
    IOUtils.copyBytes(fis, fos, configuration);
    //6.关闭资源
    IOUtils.closeStream(fis);
    IOUtils.closeStream(fos);
}
 (2)  HDFS文件下载
public static void get() throws URISyntaxException, IOException, InterruptedException {
    //1.获取配置信息
    Configuration configuration = new Configuration();
    //2.获取文件系统
    FileSystem fs = FileSystem.get(new URI("hdfs://hadoop103:9000"), configuration, "luomk");
    //3.获取输入流
    FSDataInputStream fis = fs.open(new Path("/luomk/wc1.input"));
    //4.获取输出流
    FileOutputStream fos = new FileOutputStream("/Users/luomingkui/Downloads/wc1.input");
    //5.流的对接
    IOUtils.copyBytes(fis, fos, configuration);
    //6.关闭资源
    IOUtils.closeStream(fis);
    IOUtils.closeStream(fos);
}
 
(3) 定位文件读取
    (1) 下载第一块
@Test
    public void readFileSeek1() throws IOException, InterruptedException, URISyntaxException{
        // 1 获取文件系统
        Configuration configuration = new Configuration();
        FileSystem fs = FileSystem.get(new URI("hdfs://hadoop103:9000"), configuration, "luomk");
        
        // 2 获取输入流
        FSDataInputStream fis = fs.open(new Path("/hadoop-2.7.2.tar.gz"));
        
        // 3 创建输出流
        FileOutputStream fos = new FileOutputStream(new File("/Users/luomingkui/Downloads/hadoop-2.7.2.tar.gz.part1"));
        
        // 4 流的拷贝
        byte[] buf = new byte[1024];
        
        for(int i =0 ; i < 1024 * 128; i++){
            fis.read(buf);
            fos.write(buf);
        }
        // 5关闭资源
        IOUtils.closeStream(fis);
        IOUtils.closeStream(fos);
    }
    (2) 下载第二块
@Test
    public void readFileSeek2() throws IOException, InterruptedException, URISyntaxException{
        // 1 获取文件系统
        Configuration configuration = new Configuration();
        FileSystem fs = FileSystem.get(new URI("hdfs://hadoop103:9000"), configuration, "luomk");
        
        // 2 打开输入流
        FSDataInputStream fis = fs.open(new Path("/hadoop-2.7.2.tar.gz"));
        
        // 3 定位输入数据位置
        fis.seek(1024*1024*128);
        
        // 4 创建输出流
        FileOutputStream fos = new FileOutputStream(new File("/Users/luomingkui/Downloads/hadoop-2.7.2.tar.gz.part2"));
        
        // 5 流的对拷
        IOUtils.copyBytes(fis, fos, configuration);
        
        // 6 关闭资源
        IOUtils.closeStream(fis);
        IOUtils.closeStream(fos);
    }

     (3) 合并文件
        在window命令窗口中执行
        type hadoop-2.7.2.tar.gz.part2 >> hadoop-2.7.2.tar.gz.part1