Anja, If I understand what you've asked correctly, I think you can do this with some simple SQL summarization functions. For example: select vehicle_id,count(Time_Step) from <table> group by vehicle_id; This can be accomplished with the "Extract" or "SQL Join" transformations. If your ultimate goal is to create one data set that has the data from all 20 datasets consolidated into a single dataset, rolled up by vehicle and segment, this would be a good use of the Loop transform. You would create one inner job that did the basic summarization by vehicle_id, as described above, derive the "segment name" from a parameter passed into the job (maybe the input dataset name), parameterize the name of the input table and set the options on the Table Loader for the target table to append. You would then create an outer job that calls the Inner Job using the Loop transformation, running each step sequentially instead of in parallel. If processing time is a problem, there are strategies to execute the iterations in parallel and then gather all the separate pieces later. Please see this post for a detailed explanation of how to use the Loop transform and Inner and Outer jobs in this manner. There are other ways to accomplish grouping all the datasets together, but this one will allow you to extend/reduce the number of segments without changing your job at all. Thanks, Tim Stearn
... View more