Hello, I have a process that goes to a RDBMS SQL and brings back the data from many tables by date in the following structure.
Input data structure: (all the children mentioned below can occur n number of times within a parent)
Parent1,child1, child2, child3 (child1 is child of parent1, child2 is child of child1, child3 is child of child2)
Parent2,child1, child2
Parent3,child1
Parent4,child1, child2, child3,child4
etc..
We are trying to migrate it to DataStage 7.5.2. I would like to use the Parallel server to process the data. What I would like to be able to do is build an internal structure array list of the data so that it can be traversed multiple times to process the data.
Intermediate data array:
Parent[1].child1[1].child2[1].child3
Parent[1].child1[1].child2[2].child3
Parent[1].child1[2].child2[1].child3
Parent[1].child1[2].child2[1]
Parent[1].child1[3]
Parent[2].child1
Parent[2].child2
Parent[3].child1
Parent[4].child1
Parent[4].child2
Parent[4].child3
Parent[4].child4
I can't find a way other than writing C++ routine to do this.
Has anyone come across this type of problem and solved it by using a combination of parallel stages and would like to share their finding? Or is there ways to do this in version DataStage 8?
Is there anyway to build/process arrays with using C++ ?
Moderators: chulett, rschirm, roy
Sorry but I mean to say without C++
missed the spell..posted to quickly
I mean to said with using C++
I mean to said with using C++
Sorry but I mean to say without C++
missed the spell..posted to quickly
I mean to said without using C++
I mean to said without using C++
-
- Participant
- Posts: 54607
- Joined: Wed Oct 23, 2002 10:52 pm
- Location: Sydney, Australia
- Contact:
Nothing with in DataStage allows for indexed array processing of this kind.
You could fabricate a "dynamic array" as a delimited string, but there's still no convenient way to "traverse it multiple times".
Therefore, yes, you are going to need to create some code.
You could fabricate a "dynamic array" as a delimited string, but there's still no convenient way to "traverse it multiple times".
Therefore, yes, you are going to need to create some code.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.