Yep 🙂 you can use aggregates to do this: https://opencrux.com/reference/21.04-1.16.0/queries.html#find-aggregate For example, this TPC-H uses group-by in the original SQL: https://github.com/juxt/crux/blob/9c7771e2fc40807592bd8dba65b21b1e9e18e8a2/crux-sql/test/crux/tpch_test.clj#L35-L56 And after we translated it to Datalog, see how it is now implicitly defined by use of aggregates in the :find https://github.com/juxt/crux/blob/9073b6ff85ffe49019026f41a7b0e01831da349e/crux-test/src/crux/fixtures/tpch.clj#L118-L142

Thanks, would this permit me to group docs into vectors?

Probably, although I'm struggling to think of an example(!), do you have one to hand? (feel free to DM me if it's sensitive)

If necessary you can use a predicate, subquery or custom aggregate to achieve pretty much anything