关于Java程序向Hive导入数据的问题解决方法

阅读更多

解决方法：

1、使用Linux命令，如下：

String tmpFile = "/tmp" + File.separator + TableUtil.TABLE_PREFIX_MAILING_MEMBER + "@" + mailListId + ".log";  
String scp_command = String.format("scp %s x.x.x.x:%s", fileKey, tmpFile);  
logger.info("Current TaskId#" + taskId + " : " + scp_command);  
Process process = Runtime.getRuntime().exec(scp_command);  
logger.info("Current TaskId#" + taskId + ", SCP_COMMAND : 
");  
InputStreamReader inputStream = new InputStreamReader(process.getInputStream());  
BufferedReader inputReader = new BufferedReader(inputStream);  
String line = null;  
while((line = inputReader.readLine()) != null) {  
    logger.info(line);  
}  
logger.info("Current TaskId#" + taskId + ", SCP_COMMAND : 
");  
if (inputReader != null) {  
    inputReader.close();  
}  
if (inputStream != null) {  
    inputStream.close();  
}  
logger.info("Current TaskId#" + taskId + ", SCP_COMMAND : ");  
InputStreamReader errorStream = new InputStreamReader(process.getErrorStream());  
BufferedReader errorReader = new BufferedReader(errorStream);  
line = null;  
while((line = errorReader.readLine()) != null) {  
    logger.info(line);  
}  
logger.info("Current TaskId#" + taskId + ", SCP_COMMAND : ");  
if (errorReader != null) {  
    errorReader.close();  
}  
if (errorStream != null) {  
    errorStream.close();  
}  
process.waitFor();  
logger.info("Current TaskId#" + taskId + " : " + scp_command);  
  
Thread.sleep(5000);  
// 赋权限  
String chown_command = String.format("%s '%s %s'", "ssh x.x.x.x", "chown hadoop:hadoop -R", tmpFile);  
logger.info("Current TaskId#" + taskId + " ：" + chown_command);  
process = Runtime.getRuntime().exec(chown_command);  
logger.info("Current TaskId#" + taskId + ", CHOWN_COMMAND : 
");  
inputStream = new InputStreamReader(process.getInputStream());  
inputReader = new BufferedReader(inputStream);  
line = null;  
while((line = inputReader.readLine()) != null) {  
    logger.info(line);  
}  
logger.info("Current TaskId#" + taskId + ", CHOWN_COMMAND : 
");  
if (inputReader != null) {  
    inputReader.close();  
}  
if (inputStream != null) {  
    inputStream.close();  
}  
logger.info("Current TaskId#" + taskId + ", CHOWN_COMMAND : ");  
errorStream = new InputStreamReader(process.getErrorStream());  
errorReader = new BufferedReader(errorStream);  
line = null;  
while((line = errorReader.readLine()) != null) {  
    logger.info(line);  
}  
logger.info("Current TaskId#" + taskId + ", CHOWN_COMMAND : ");  
if (errorReader != null) {  
    errorReader.close();  
}  
if (errorStream != null) {  
    errorStream.close();  
}  
process.waitFor();  
logger.info("Current TaskId#" + taskId + " : " + chown_command);   

 2、使用如下方法：

 

public static void copyLocalFileHdfs(String localPath, String hdfsPath) throws IOException {
    Configuration conf = new Configuration();
    conf.set("fs.defaultFS", "hdfs://Ucluster");
    conf.set("dfs.nameservices", "Ucluster");
    conf.set("dfs.ha.namenodes.Ucluster", "xx,yy");
    conf.set("dfs.namenode.rpc-address.Ucluster.xx","hostname:port");
    conf.set("dfs.namenode.rpc-address.Ucluster.yy","hostname:port");
    conf.set("dfs.client.failover.proxy.provider.Ucluster",
            "org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider");
    try {
        FileSystem fs = FileSystem.get(conf);
        Path src = new Path(localPath);
        Path dst = new Path(hdfsPath);
        fs.copyFromLocalFile(src, dst);
    } catch (Exception e) {
        e.printStackTrace();
    }
}