Well, for the case presented by the OT the cartesian join does not look sensible. So it at least requires an explanation why this should be like this.
Combining tables based on no common field or condition (the cross join/cartesian join) is something that does not make sense in a normalized data model.
Using cartesian joins to blow up result sets is a technique that does have it's limitation. For use cases like disaggregation - like the one you mentioned - SAP HANAs planning engine uses more efficient algorithms than incorporating it into the SQL statement.
Similarly for time series, there are procedures available for disaggregation of time series data.
Hints, and this is my personal opinion, are often workarounds that should help out when the original problem is located somewhere else (e.g. misunderstanding of the data, the way SQL works, or what the used DB engine does with queries). And when applied, those hints are often poorly understood and cannot later on be managed properly (e.g. when the data or the DB engine changes).
Coming back to your case I think it does not make sense to mix up your situation and the one presented by the OP. So, rather open a thread on your own, providing the information required to understand what's happening in your case.
- Lars