In Oracle I can declare a reference cursor...
TYPE t_spool IS REF CURSOR RETURN spool%ROWTYPE;
...and use it to pass a cursor as the return value...
FUNCTION end_spool
RETURN t_spool
AS
v_spool t_spool;
BEGIN
COMMIT;
OPEN v_spool FOR
SELECT
*
FROM
spool
WHERE
key = g_spool_key
ORDER BY
seq;
RETURN v_spool;
END end_spool;
...and then capture it as a result set using JDBC...
private Connection conn;
private CallableStatement stmt;
private OracleResultSet rset;
[...clip...]
stmt = conn.prepareCall("{ ? = call " + call + "}");
stmt.registerOutParameter(1, OracleTypes.CURSOR);
stmt.execute();
rset = (OracleResultSet)stmt.getObject(1);
What is the equivalent in MySQL?
Mysql has an implicit cursor that you can magically return from a stored procedure if you issue a select.
Here's an example:
CREATE PROCEDURE `TEST`()
MODIFIES SQL DATA
BEGIN
SELECT * FROM test_table;
END;
and in your java code:
String query = "{CALL TEST()}";
CallableStatement cs = con.prepareCall(query,
ResultSet.TYPE_SCROLL_INSENSITIVE,
ResultSet.CONCUR_READ_ONLY);
ResultSet rs = cs.executeQuery();
Googling on cursors in MySQL, it doesn't seem like you can actually return a Cursor from a Proc or Function. Additionally, I found the following in the MySQL JDBC manual:
"MySQL does not support SQL cursors, and the JDBC driver doesn't emulate them, so "setCursorName()" has no effect."
In general, I believe Oracle's implementation here breaks JDBC, and is not used elsewhere (MySQL, MSSQL, etc). You should be returning your results as a select statement and iterating over the JDBC ResultSet, as is standard (and intended) practice when using JDBC.
fill a temporary table in a procedure and just read the temporary table... :)