Question entered earlier: still not showing up on PB forum
I am trying to determine if setting the transaction isolation level will help me resolve a problem that my customers are having. I am working with an application window with several tabs. The window has 9 tabs and they update various tables. An update to tab 2 through 9 might affect tab 1 because tab 1 is the General tab . It summarizes the other tabs in addition to having unique information
Problem: Occasionally two customers open the same claim. One customer is updating tab- x. Another is updating a different tab- y. The window was written so that when a customer hits the <update> button, the update function cycles through all of the tabs and saves to the appropriate tables. Both start with the same data. However, whoever saves last ends up clobbering the prior customer’s updates. If customer 2 is last and updates tab y, s/he still has an old copy of what was on tab x.
I tried setting the transaction isolation level to RC thinking that would force a phantom read before the second update, but that doesn’t appear to happen. How does setting the transaction isolation level affect powerbuilder updates? And if the update is actually done in a stored procedure, does setting the transaction isolation level in PB mean anything?
Is there another way to do this? I am preparing to implement a clunky (imho) option of comparing the “last updated timestamp” to the “open for update timestamp” and telling the customer that she needs to refresh because someone else has updated her record.
Environrment : PB 12.1
Db: oracle 9i