DSXchange

I think aggregator can only be used when we have standard values for example
if we are expecting all 1 s and 0 s as data values for these columns then i think its pretty straight forward (as you and others have mentioned ) but the if we get values like 'abc' , 'def' so on and so forth then I am guessing the stage variables and retaining the last row of the group logic (provided by james would work ) .

I tested both ways successfully but I am assuming using stage variables will be "best value" as you mentioned James.

Thanks all . I am marking this issue as resolved.

So long as you understand that the approach I was suggesting is simply aimed at keeping a best or desired value, when present, over all others in the key group. That's what my use of the term "best value" means in this example.

Regards,

I would aim for the most straight-forward approach that fits the requirements. If all you need is the Aggregator, use it. If you need to accomodate other values, then take a more 'complex' approach.