spark sql完成jdbc数据源导入hive表，java API实现

1、基础环境

spark on yarn环境已经搭建，通过start-thriftserver.sh --master yarn启动 sparksubmit进程即可

dm数据库；

2、使用jdbc来访问spark，java API实现

2.1 创建datasource、connection、statement代码，这里属于通用代码，使用Druid连接池来创建jdbc连接

 /**
     * 读取druid.properties,并生成datasource
     */
    private static Statement init() throws Exception {
        Properties properties = new Properties();
        InputStream resourceAsStream =
            DataHubTask.class.getClassLoader().getResourceAsStream("conf/druid.properties");
        try {
            properties.load(resourceAsStream);
        } catch (IOException e) {
            LOG.error("The resource loaded failed, please check the path {}.", SPARK_THRIFT_SERVER_PATH);
            throw new RuntimeException(e);
        }

        dataSource = DruidDataSourceFactory.createDataSource(properties);
         
        Connection connection = dataSource.getConnection();