使用Java / JDBC的Hive表大小

Question

Is it possible to find the sizes in GB of the hive tables using Java/jdbc? I don't want to depend on the hive warehouse folder in HDFS (as described in link ) as different tables may have different locations

Answer 1

Currently not be possible but it can be done .

To get file size you have to run on file system( HDFS) commands .

In case of RDMS data bases ie sql server have encapsulated file system commands in SYS views and SYS functions (DMF and DMVs) .

If some one write or develop such UDF it will be possible but Internally the UDF will be calling same comand.

Answer 2

If you are mention 'totalSize' from 'tblproperties' then it is possible with similar approach:

String driverName = "org.apache.hive.jdbc.HiveDriver";
String connectionURL = "jdbc:hive2://HOSTNAME:PORT/default";

try {
     Class.forName(driverName);
     Connection connection = DriverManager.getConnection( connectionURL, "", "");
    } catch (ClassNotFoundException e) {
      e.printStackTrace();
      System.exit(1);
    }

Statement stmt = connection.createStatement("show tblproperties TABLENAME");
ResultSet rs =  stmt.executeQuery(stmt);
while(rs.next()){
            //doWhatYouWant
        }