|
Home > Archive > BizTalk Server Orchestration > April 2004 > Convoy Scenario
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
|
|
|
| Hi, I am trying to achieve a convoy in BTS 2004, but having some trouble. I
understand the convoy pattern, but I can't find any good examples for what I
need. I am trying to achieve a "Uniform Sequential Convoy", but I have no
idea what expression to put in my loop shape. Am I missing something
obvious?
My scenario...
A large batch of xml messages are dumped into a directory at the same time,
half of them have one customer ID (say 1) and half of them with customer ID
2. What I want to be able to achieve is two orchestration instances running
for the files, one processing all messages for customer ID 1 and the other
processing all customer messages for customer ID 2. This is important
because I am trying to avoid a race condition. I know that I can correlate
on customer ID, and have managed to get two messages correlating fine.
How do I get this working when say 1000 messages arrive at the same time for
one customer, and the messages need to be processed one by one? What
expression do I need to put in my loop shape to get this type of convoy
working?
| |
| Matt Milner 2004-04-14, 7:37 pm |
| You have to figure out when to stop basically, so you could either base that
on time or number of messages.
If you know how many messages, which it doesn't seem you do, then you could
increment a counter on each time through the loop and stop when you had
processed all the messages.
Otherwise, you might but a datetime variable in that you set on the first
message. Then check in the loop whether a certain amount of time has passed
since you started processing and quit when you reach that threshold.
Matt
"Danny" <danny.lee@nospamsdm.com.auuuu> wrote in message
news:eHWorJfIEHA.3356@TK2MSFTNGP11.phx.gbl...
> Hi, I am trying to achieve a convoy in BTS 2004, but having some trouble.
I
> understand the convoy pattern, but I can't find any good examples for what
I
> need. I am trying to achieve a "Uniform Sequential Convoy", but I have no
> idea what expression to put in my loop shape. Am I missing something
> obvious?
>
> My scenario...
>
> A large batch of xml messages are dumped into a directory at the same
time,
> half of them have one customer ID (say 1) and half of them with customer
ID
> 2. What I want to be able to achieve is two orchestration instances
running
> for the files, one processing all messages for customer ID 1 and the other
> processing all customer messages for customer ID 2. This is important
> because I am trying to avoid a race condition. I know that I can
correlate
> on customer ID, and have managed to get two messages correlating fine.
>
> How do I get this working when say 1000 messages arrive at the same time
for
> one customer, and the messages need to be processed one by one? What
> expression do I need to put in my loop shape to get this type of convoy
> working?
>
>
|
|
|
|
|