I could do it with dbt but it would be much more efficient to do it as part of the extraction process. Renaming the destination stream based on a bit of data in the record, before it gets to the target, is a trivial amount of work. Scanning the table N times after its been landed into the destination is N-1 times more expensive. 🤷🏻♂️ Not a blocker or anything, just trying to be efficient. 😉