Group Determination

[vinodn]

I need help regarding group ditermination.

My input looks like this. the first 3 columns are my key.

10,20,30,A
10,20,30,X
10,20,40,B
10,20,50,Y
10,20,50,Z
10,20,60,D

My output need to be

1st o/p :

10,20,30,A
10,20,30,X
10,20,50,Y
10,20,50,Z

2nd o/p
10,20,40,B
10,20,60,D

I can do this by using the count function in the aggregator and where count >1 (by grouping on the 1st 3 columns) can be routed to 1st output and count =1 to the 2nd output, I want to know if this can be done in the transformer(like the dedup logic).

Thanks for your help.

[ray.wurlod]

If you're in a version of DataStage that supports looping in the Transformer stage then yes. Otherwise probably not.

[qt_ky]

I would think you can by using the Sort stage key change column plus Transformer stage.