Thread: JDBC CTS failures

There is a test case in the JDBCCTS that uses the following


CallableStatement.prepareStatement( call function that does updates )
then does a number of addBatch callls
and does executeBatch()

This fails since executeBatch does not allow resultsets and we
emulate calling a function by doing "select * from funcname()" which
does return a result set.

The proposal is to have CallableStatement.executeBatch ignore this
resultSet instead of throwing an exception.

Can anyone see any problems with this ?

Dave

On Thu, 19 Oct 2006, Dave Cramer wrote:

> CallableStatement.prepareStatement( call function that does updates )
> then does a number of addBatch callls
> and does executeBatch()
>
> This fails since executeBatch does not allow resultsets and we emulate
> calling a function by doing "select * from funcname()" which does return a
> result set.
>
> The proposal is to have CallableStatement.executeBatch ignore this resultSet
> instead of throwing an exception.
>
> Can anyone see any problems with this ?
>

There are potential problems if you pass large parameters to the function
and it returns large values.  See the comments around
org.postgresql.core.v3.QueryExecutorImpl.MAX_BUFFERED_QUERIES.
Despite the potential problems I think we should implement this
solution because we've gotten this complaint reasonably often and the
errors should be rare (although unfortunately difficult to diagnose).

Kris Jurka