Hi Ann,

thanks for confirming it.

> >- for write transaction
> >"write, nowait, rec_version, read_committed"
> >
> For write transactions, I prefer wait as it reduces the chances of
> certain kinds of performance problems.

Could you explain this a little bit more, please .. In my
application each user can see all records and add records, but
modify only the records created by himself - in this case the wait
lock shouldn't be possible?

best regards,
Harri