>
>Actually what I do is catch the update conflict and then just wait a short
>time and retry the isc_dsql_execute. This is better than the rollback I
>think and should work properly in all cases. It sort of makes it work how I
>wanted it to except that I have to do the loop and retries instead of the
>database doing it.

>Thanks for your response. It really cleared up a lot for me and I think I've
>got it all working how I want now!

I have no idea why that works - in theory a transaction sees a stable
view of the database so if there's a conflict once, there'll be a conflict
next time. However, if it works, why not?

Note that InterBase was designed to support multiple concurrent
transactions from a single connection - allowing you to encapsulate
an operation like this. None of the middle layers (except IBO?)
support that capability... We also created generators to avoid
exactly this problem... again, not portable.


Regards,

Ann
www.ibphoenix.com
We have answers.