T
Thomas Mercer-Hursh
Guest
And, in many applications, even if the record was updated in between, one would want the new version. Only if there were a meaningful time delay ... e.g., if it went out to the user in between, would one worry and then, of course, one would use standard optimistic locking behavior to detect the possible change. The delete in between should be catchable and thus gracefully handled. Note that doing the second find with a GUID or whatever doesn't actually change the scenario except for your third case, which seems a fairly thin chance. Which said, I can't see any reason why a find by RECID would be better than a find by GUID unless one was dealing with legacy code where one had to use multiple fields to uniquely identify the record.
Continue reading...
Continue reading...