|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
 |
|
Aggregation without intermediate output |
« View previous topic :: View next topic » |
Author |
Message
|
SpitFire |
Posted: Fri Apr 14, 2006 12:32 am Post subject: Aggregation without intermediate output |
|
|
 Acolyte
Joined: 01 Aug 2005 Posts: 59 Location: India
|
Hi all,
I'm using WBIMB 5.0 and I'm trying a flow with aggregate (control, request & reply) nodes. The rough sketch of the flow is basically as follows:
(1) MQ In
(2) Agg Ctrl <--- from(1)
(3.1) Compute-1 <-- from (2)
(3.2) Compute-2 <-- from (2)
(4.1) Agg Req-1 <-- from (3.1)
(4.2) Agg Req-2 <-- from (3.2)
(5) Agg Reply <-- from (4.1 & 4.2)
(6) Compute-3 <-- from (5)
(7) MQ Out <-- from (6)
The aggregation is done only to do some processings on the msg in 'parallel'. Can I implement it like this? There is no intermediate output where the messages are sent and replies are collected. The processing happens at Compute (Compute-1 & Compute-2) node itself.
However, when I try this I get the following error in 'ExceptionList'.
[IBM][CLI Driver][DB2/NT] SQL0803N One or more values in the INSERT statement, UPDATE statement, or foreign key update caused by a DELETE statement are not valid because the primary key, unique constraint or unique index identified by "1" constrains table "MQSIUID.BAGGREGATE" from having duplicate rows for those columns. SQLSTATE=23505
What is that I'm overlooking? _________________ ...: 5|71+ph1|23 :... |
|
Back to top |
|
 |
jefflowrey |
Posted: Fri Apr 14, 2006 2:46 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002 Posts: 19981
|
You're not going to add any parallelism this way, unless you push Compute-3 into another flow.
Think of each node in your flow as a subroutine call. _________________ I am *not* the model of the modern major general. |
|
Back to top |
|
 |
SpitFire |
Posted: Fri Apr 14, 2006 3:10 am Post subject: |
|
|
 Acolyte
Joined: 01 Aug 2005 Posts: 59 Location: India
|
jefflowrey wrote: |
You're not going to add any parallelism this way, unless you push Compute-3 into another flow.
|
Agreed. That's the plan I'm eventually going to adopt.
However, I'm getting stuck when I don't use an MQOutput node after the aggregate control node.
I tried populating the 'WrittenDestination' tree with appropriate values, that are required by the 'AggregateRequest' node. Still, I'm not able to succeed. _________________ ...: 5|71+ph1|23 :... |
|
Back to top |
|
 |
wschutz |
Posted: Fri Apr 14, 2006 3:24 am Post subject: |
|
|
 Jedi Knight
Joined: 02 Jun 2005 Posts: 3316 Location: IBM (retired)
|
Are you saying you want Compute1 and Compute2 to run in parallel? Why? _________________ -wayne |
|
Back to top |
|
 |
SpitFire |
Posted: Fri Apr 14, 2006 3:36 am Post subject: |
|
|
 Acolyte
Joined: 01 Aug 2005 Posts: 59 Location: India
|
wschutz wrote: |
Are you saying you want Compute1 and Compute2 to run in parallel? Why? |
To reduce computation time, of course. Is that not possible? Can't they be run in parallel, say as jefflowrey quoted, I'm also designing appropriate subflows after the aggregate control node to ensure the computation inside the compute nodes run in parallel fashion? _________________ ...: 5|71+ph1|23 :... |
|
Back to top |
|
 |
|
|
 |
|
Page 1 of 1 |
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
|
|
|