Aggregator Scenario

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
vijayrc
Participant
Posts: 197
Joined: Sun Apr 02, 2006 10:31 am
Location: NJ

Aggregator Scenario

Post by vijayrc »

Hi,
My Input Fields are FLD-1,FLD-2,FLD-3,FLD-4,FLD5,FLD-6,FLD-7
AGG1 Groups on FLD-1 sums on FLD-4 & FLD-5
AGG2 Groups on FLD-1,FLD-2 sums on FLD-4 & FLD-5
AGG3 Groups on FLD-1,FLD-2,FLD-3 sums on FLD-4,FLD-5,FL7
Can I FUNNEL the AGG1,AGG2,AGG3 output into one as Output File layout [similar to Input File layout]

...........--->AGG1--->
Seq IN -->AGG2--->Funnel---> Seq Out
...........-->AGG3--->

Above isn't possible, as the meta data of all input data sets must be identical. So what other inexpensive, simple STAGE after each AGGR and before FUNNEL can I use to get the varying fields from diff AGGR into one common layout for the funnel??.

Thanks
Vijay
Last edited by vijayrc on Tue Oct 10, 2006 5:37 pm, edited 1 time in total.
kris007
Charter Member
Charter Member
Posts: 1102
Joined: Tue Jan 24, 2006 5:38 pm
Location: Riverside, RI

Post by kris007 »

The inputs to the Funnel stage needs to have identical(exact) metadata. So, your job desing won't work unless you are bringing out all the columns from all the aggregators which I doubt it. You could use a join stage or Merge stage followed by a Transformer.
Kris

Where's the "Any" key?-Homer Simpson
vijayrc
Participant
Posts: 197
Joined: Sun Apr 02, 2006 10:31 am
Location: NJ

Post by vijayrc »

kris007 wrote:The inputs to the Funnel stage needs to have identical(exact) metadata. So, your job desing won't work unless you are bringing out all the columns from all the aggregators which I doubt it. You could use a join stage or Merge stage followed by a Transformer.
Thanks. I have bunch of Aggregators, and I don't want to have a transformer[for performance issue] for each Aggregator to get to the same metadata for the funnel. Any other Stage, that's inexpensive, and simple to get to a solution for this ??
[1]Would a Colum Generator take the AGGR output fields, and initialize the other fields for the funnel input ?? I'm more inclined towards this
[2] Can I use a COPY stage ?? Would this work and be simple
ray.wurlod
Participant
Posts: 54607
Joined: Wed Oct 23, 2002 10:52 pm
Location: Sydney, Australia
Contact:

Post by ray.wurlod »

Copy stage can not add columns to the record schema. Column Generator stage can.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
kris007
Charter Member
Charter Member
Posts: 1102
Joined: Tue Jan 24, 2006 5:38 pm
Location: Riverside, RI

Post by kris007 »

vijayrc wrote: I have bunch of Aggregators, and I don't want to have a transformer[for performance issue] for each Aggregator to get to the same metadata for the funnel. Any other Stage, that's inexpensive, and simple to get to a solution for this ??
[1]Would a Colum Generator take the AGGR output fields, and initialize the other fields for the funnel input ?? I'm more inclined towards this
From what I understand about that stages, it should work fine. I have never tried it though.
vijayrc wrote: [2] Can I use a COPY stage ?? Would this work and be simple
I dont think it is possible.
Kris

Where's the "Any" key?-Homer Simpson
pigpen
Participant
Posts: 38
Joined: Thu Jul 13, 2006 2:51 am

Post by pigpen »

What about that? Add transformer after Agg and add dummy columns inside to sync all inputs before the Funl.
vijayrc
Participant
Posts: 197
Joined: Sun Apr 02, 2006 10:31 am
Location: NJ

Post by vijayrc »

pigpen wrote:What about that? Add transformer after Agg and add dummy columns inside to sync all inputs before the Funl.
Thanks Kris and Ray..and I would try out the Col Gen, though I was skeptical abt the COPY stage.

Pigpen, I have bunch of Aggregators[20+] for different purposes, and don't want to burn myself out w/ a Transformer in each line just to get those output fields, with no value in them. Thanks anyways
Post Reply