A possible way is: use the Aggregator to aggregate the features grouping by ColB and also create a list attribute storing ColA values, find '2' in the list with the ListSearcher, and restore the original features from only the NotFound features with the ListExploder.

Hi @takashi. Maybe, I had put the statement in a partial way. I'll explain it a bit.

I have roads like shown below.

There are groups of roads that I have selected using SQLCreator from the server. Now, I have to find the specific group(like Colb(above), so all the road segments have a common group name) of roads(for which you told the above solution). And after that, I need to find how many roads are just one way and calculating their percentage(info not displayed here).

So, if I aggregate the values like above mentioned, calculating percentage would be a bit of task.

Also, I tried reading about creating list attribute transformers, but couldn't get much.

Thanks.

You could use this sort of flow, test for value of 2, then use any features that pass as suppliers in a featuremerger. If you keep only unmerged features that 'discards' the whole group