JDBC Mysql Resultset too large!

Question

Hello everybody,
I'm currently building a large database retrieval system(JAVA and mysql). Everything works fine, but I have a big problem. The returning resultset is way higher then it should be!!
The table has the following structure:
Probe_set_id unsigned mediumint(6)
Expressionvalue float(5,1)
Dataset unsigned tinyint(2)
I'm trying to retrieve a dataset with 6.5 million rows with the following query.
Select Probe_Set_ID, Expressionvalue from Test5 where dataset=1
This results in JAVA taking around 500 megabyte retrieving the Resultset!!! If my calculations are right it should take around the following:
Float 4 bytes * 6500000 = 26000000 bytes
unsigned mediumint = long 8 bytes * 6500000 = 52000000 bytes
That would be 78000000 bytes, which is around 75 megabyte and not 500 megabyte!!!!!
Code sample:
               String dataBaseName = "testDB";
               String dataBaseUser = "java";
               String dataBasePass = "clustering";
               String url = "jdbc:mysql://localhost/" + dataBaseName;
               Class.forName ("org.gjt.mm.mysql.Driver").newInstance ();
               Connection conn = DriverManager.getConnection (url, dataBaseUser, dataBasePass);
               conn.setReadOnly(true);
               Statement s = conn.createStatement();
               s.execute("Select Probe_Set_ID, Expressionvalue from Test5 where dataset=1");
               ResultSet rs = s.getResultSet();
This results in a giant resultset and I can't see the answer. I hope someone can help me. I'm using the newest JDBC driver of mysql!

Answer 1

Yes the JAVA application really takes up this amount! It is really akward. The DB doesn't take a real big amount of memory because of the buffer sizes I've set. I execute the java application as follow:
./java -server -XX:MaxHeapFreeRatio=5 -XX:MinHeapFreeRatio=20 -Xmx1512M dbTest
I'm really trying to keep the heap small because the JAVA application ( a gene expression processing application/server) takes big amounts of memmory when downloading data from the mysql database and the server is accesible by multiple users parallel. Though I'm still testing alone, so the memmory abuse is not a cause by other users.