As I understand it, you need to aggregate and group by the account in any case. The question is whether the aggregation returns 1 group (i.e. when the account is unique) or more than 1 group (i.e. when the account is not unique).
In this case, you can calculate a column with the total number of accounts, and then do conditional branching depending on whether you have 1 or more accounts in your dataset.
See below an example that does that:
non-unique-branching.morph (4.4 KB)
