Aggregator - grouping based on all columns

[vnspn]

Hi,

We have some kind of uncertainty when we are trying to implement a Server job's Aggregation logic to a Parallel job.

We have 4 columns - A, B, C, D; as input to the Aggregator. The grouping keys in the Aggregator is all the 4 columns. The output of the Aggregator is also the same 4 columns.

Our requirement is to Group by all the 4 columns to get the unique combination of the 4 columns put-together.

Please suggest the way this be done using the Aggregator stage in a parallel. job. It was previously done using the Aggregator stage in the server job. In the parallel job's Aggregator stage, when the Aggregation type is selected as 'Calculation', it requires to add properties to it like the Min, Max, etc. Thats where we are stuck.

Thanks.

[ray.wurlod]

Upstream of the Aggregator stage use a Column Generator stage to generate a new column containing a constant. Perform the calculation on that. Discard it downstream of the Aggregator stage, perhaps in a Copy stage (which is permitted to drop columns).

[vnspn]

Thanks Ray.

I would try this method.