I haven't yet used this option, but I believe this is how it works. I don't have context for what the data look like, but I assume you understand that the "many-to-many" relationship is not default for a reason, which implies one should carefully consider if that really should be happening and what the consequence of it is.

Yes indeed.

There are several kinds of joins here in which we require this type of many-to-many mapping.

1. Joining scenarios to ABCD: There are many scenarios that map to many companies, we expect the mapping to not be distinct
2. Joins to region_isos: There are likely to be many companies with the same plant location ISO, and many regions containing that plant location ISO, we expect the mapping to not be distinct
3. Joins to green_techs_in_sector: We expect the same sector could pop up multiple times on the left-hand-side, and we expect there to be multiple green techs for any given sector

I believe all of these are expected.