hi @Abby 

The group aggregator has two output endpoints. 

• out_results endpoint contains the result of your group aggregator step
• out sends the input records to subsequent steps.

You can use a join step after the group aggregator, it joins two separate data flows into a single data flow, this is the way to keep all columns 

Hi ​@alexAguilarMx based on your suggestion, I applied the same to my use case but I found out that the left and the right join produces the same column names. For example, SUPPLY_CREATE_DATE that is created in one of the aggregate sets appear twice both on left and right after the join, so I have left.SUPPLY_CREATE_DATE and right.SUPPLY_CREATE_DATE. Ideally, after the join, I should take the right side for the aggregation since I would expect the column to come from the out_results only?

On a side not, I have several aggregator sets which each set has condition on the ‘when’ configuration.