Kristian Waagan-4 wrote:
>
> wfalby wrote:
>> My application processes events. These events are sent to registered
>> users.
>> These events can be deleted when they reach a certain age. There is a
>> thread
>> to send events to registered users and there is another thread to delete
>> old
>> events. The thread sending an event should lock it from the thread that
>> deletes old events. Each thread uses its own connection to the database.
>> In
>> the sending thread, I've tried using a SELECT ... FOR UPDATE and setting
>> the
>> ResultSet.CONCUR_UPDATABLE on the prepare statement. In both cases, the
>> thread that deletes events deleted the event being held by the sending
>> thread. The default isolation level and autocommit options are used.
>>
>
> Hi Walter,
>
> If you're using the read committed isolation level, I think you need to
> use SELECT ... FOR UPDATE WITH RS.
> Which behavior do you see if you do that?
>
> I think Derby treats SELECT ... FOR UPDATE a bit differently than some
> other database systems, and I also believe there is at least one Jira
> [1] issue logged for changing the behavior.
>
>
> Regards,
> --
> Kristian
>
> [1] https://issues.apache.org/jira/browse/DERBY
>> Thanks in advance...Walter
>>
>
>
>
Kristian:
Thanks for the info. I tried your suggestion, but all rows are deleted. I'll
check the Jira issue to see if it fits my situation.
Walter
--
View this message in context:
http://www.nabble.com/Row-Locking-tp23333804p23366698.html
Sent from the Apache Derby Users mailing list archive at Nabble.com.
