On 1/14/2011 12:43 PM, Aldo Caruso wrote:

> But it seems to me nearly impossible that another transaction could
> delete the record in that moment ( because of the business logic, those
> record aren't deleted until the end of the day ). But I'll continue to
> investigate ...

Sorry, I was thinking of the way the database works internally. An old
record version is deleted when a newer version is committed and visible
to all running transactions. The operation that created the new version
could be a delete, leaving a deleted stub, or it could be an update. So
you'd see that behavior after a committed update.


Cheers,

Ann