Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
![]() |
![]() |
Home -> Community -> Usenet -> c.d.o.misc -> Re: Allowing Oracle transactions to span client processes
> In my mind it means that app A makes some changes to table FOO. And
> commit. However this commit must not change the live FOO table. App B
> comes along and picks up where app A has left. Etc. Until app D
> decides enough is enough and commit these changes tp the live FOO
> table. Each of these transactions in the workspace behaves like a
> normal second/sub-second transaction - it simply does not change the
> data in LIVE. It works within the current selected workspace version
> of that data.
Yeah, that's what I want.
> That to me means using Oracle Workspace.
Ok. I'll need to have another look at the Workspace docs I think.
> And yes - the clients will need to know in what workspace to work. But
> that is an implementation issue and IMO easily solved (using anything
> from a PL/SQL calls to a database session trigger).
So each of my sessions could run some PLSQL that would effectively 'enter' and 'exit' a given workspace? And I can then commit or roll the whole workspace back using another PLSQL proc? If this is the case then maybe I can use it - I was thinking that Oracle Workspace had a GUI or admin tool of its own with which I would have to define a transaction workspace before any session can use it. Then I started to worry that I'd need to have a Java/C++ API top call into some Workspace sub-system or something... :(
Thanks a lot for your help... I'll look into Workspaces more deeply.
Cheers
Ben
Received on Mon Sep 22 2003 - 04:27:07 CDT
![]() |
![]() |