| Author |
Message
|
| stiffann |
 Posted: Tue Jul 02, 2002 1:31 am Post subject: How do I define global variables |
 |
|
Apprentice
Joined: 17 Mar 2002
Posts: 26
|
Hi all,
I would like to define a global variable, a counter for instance, that would be accessible by all my message flows deployed in a particular Execution Group.
The aim of it is to increment a counter each time a message flow is invoked, and to trigger a specific action when the counter has a specific value.
Thank you in advance !!  |
|
| Back to top |
|
 |
| nico_fr |
| Posted: Tue Jul 02, 2002 1:53 am Post subject: Environement variables |
|
|
Newbie
Joined: 29 May 2002
Posts: 2
Location: France
|
In WMQI 2.1 you can save your values in the "Environement" variable.
ex:
"SET Environment.Variables.MyError='y error';"
Consequenly you'll have in memory during all the messageflow the value:
${Environment} :
(
(0x1000000)Variables = (
(0x3000000)MyError = 'My error'
) |
|
| Back to top |
|
|
| stiffann |
| Posted: Tue Jul 02, 2002 2:15 am Post subject: |
|
|
Apprentice
Joined: 17 Mar 2002
Posts: 26
|
The official IBM documentation (Working with Messages) says that "an empty Environment tree is created when a message is received by the input node and parsed. It has no structure or content until you populate it using ESQL statements".
So, that's why I thought that the variables created were only visible within the concerned message flow and for the concerned message.
I'll try some experimentations to verify your information.
Thanks ! |
|
| Back to top |
|
|
| stiffann |
| Posted: Tue Jul 02, 2002 2:54 am Post subject: |
|
|
Apprentice
Joined: 17 Mar 2002
Posts: 26
|
As I said before, and after having experimented by myself, it seems that the "Environment" tree is created for each new message. It means that I can't store global variables in this tree.
If someone else has a good idea for that, I'm still listening to you ! 
Thanks in advance |
|
| Back to top |
|
|
| CodeCraft |
| Posted: Tue Jul 02, 2002 3:53 am Post subject: |
|
|
Disciple
Joined: 05 Sep 2001
Posts: 195
|
Have a check around the IBM Support Packs. I think I heard before of someone writing a node that had shared memory handling for variables between flows.
The alternative, a two column one row table in a database:
Create Table Counter
( MyKey Char(15) Primary Key,
MyValue Integer
)
You might be able to do some tuning to ensure that table remains cached. Can't think of anything nicer right now. |
|
| Back to top |
|
|
|
|
