17.03.2012 14:03, ainpoissee wrote:

>
> Queries which use different connections obviously run in parralel, but
> I'm wondering exactly how it works when two queries share an connection.
> Are they serialised so that second one won't run until first one is done
> (ie second one would be sent to the server only after the full resultset
> of the first one is fetched)?

They're serialized at the API call level. So the second one won't be
necessarily waiting for the whole result set to be fetched, it could
start working just after the first fetch call.


Dmitry