how to send different partions data to different targets?

Post questions here relative to DataStage Enterprise/PX Edition for such areas as Parallel job design, Parallel datasets, BuildOps, Wrappers, etc.

Moderators: chulett, rschirm, roy

Post Reply
jhansi
Participant
Posts: 33
Joined: Sat May 26, 2012 4:56 am

how to send different partions data to different targets?

Post by jhansi »

i hav data like this
1
1
2
2
2
3
how to send all 1's to one target,all 2's to other target and all 3's to another target.
can i use partition tech?
Last edited by jhansi on Thu Jun 14, 2012 2:26 am, edited 1 time in total.
jhansi
stuartjvnorton
Participant
Posts: 527
Joined: Thu Apr 19, 2007 1:25 am
Location: Melbourne

Post by stuartjvnorton »

Why not just use a filter?
SURA
Premium Member
Premium Member
Posts: 1229
Joined: Sat Jul 14, 2007 5:16 am
Location: Sydney

Re: how to send different partions data to different targets

Post by SURA »

If you have 4................400
Thanks
Ram
----------------------------------
Revealing your ignorance is fine, because you get a chance to learn.
ray.wurlod
Participant
Posts: 54607
Joined: Wed Oct 23, 2002 10:52 pm
Location: Sydney, Australia
Contact:

Post by ray.wurlod »

Welcome aboard.

When's the interview?

Why the requirement to use partitioning technology?

It probably can be done, but would not be for the faint-hearted. As SURA suggests, the partitioning technology handles from one to a countably infinite number of processing nodes, so you would need to devise a naming strategy for your targets.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
jhansi
Participant
Posts: 33
Joined: Sat May 26, 2012 4:56 am

Post by jhansi »

stuartjvnorton wrote:Why not just use a filter?
i can use filter or transformer by giving contraint like
DSLink5.sno =1 Then DSLink5.sno Else 0

.but what if i hav large amount of data?
jhansi
BI-RMA
Premium Member
Premium Member
Posts: 463
Joined: Sun Nov 01, 2009 3:55 pm
Location: Hamburg

Post by BI-RMA »

I don't see how you are going to accomplish something that is going to give you a dynamic number of output links from whatever stage you might use.

To use partitioning techniques you would need to know how many keys of the type you provided you are going to get and the number of rows per key should be at least almost equal, because otherwise your partitioning will be inefficient due to skewing.

If you know which key values are to be expected, you can of course use transformer-constraints to split your data. You can also define an "Otherwise"-link, that will take all the data you did not provide specific links for.
"It is not the lucky ones are grateful.
There are the grateful those are happy." Francis Bacon
SURA
Premium Member
Premium Member
Posts: 1229
Joined: Sat Jul 14, 2007 5:16 am
Location: Sydney

Post by SURA »

jhansi wrote:but what if i hav large amount of data?
Please understand clearly what the requirement is, and then start to think about the logic. Think max possibilities and ask questions with whoever given this task. Then start looking for solution!
Thanks
Ram
----------------------------------
Revealing your ignorance is fine, because you get a chance to learn.
stuartjvnorton
Participant
Posts: 527
Joined: Thu Apr 19, 2007 1:25 am
Location: Melbourne

Post by stuartjvnorton »

jhansi wrote:
stuartjvnorton wrote:Why not just use a filter?
i can use filter or transformer by giving contraint like
DSLink5.sno =1 Then DSLink5.sno Else 0

.but what if i hav large amount of data?
Think about it for a second.
When you partition, you can use a number of methods to work out which partition to put a record in.

In your case, you specified integer values. Using a modulus to split the output would be trivial.
Post Reply