The ability to join collections in real-time really allows developers to build complex use cases. Currently, I am having to denormalize data so excessively that the solution doesn't seem to scale. Having to implement workaround after workaround to circumvent the limitations.

An aggregate lookup function similar to that of mongodb would really be a life save. We are ok to compromise on the extra reads, and some performance if it means to bring us data consistency.

Really having this feature would stop people looking into RDBMS based alternative such as Supabase.