Let’s say in my Foundry enrollment I have two organizations - A and B. I also have 3 projects:
Now let’s say in Project C I join datasets from Project A and Project B into a single output dataset. Would a user looking at the output dataset need either or both organizations A and B to be able to see the data?
You will need both organizations A and B to be able to see the data. Otherwise, you will need to stop_requiring the markings of both inputs to the transformation that joins the datasets into Project C. This will allow the output dataset to inherit the disjunctive markings of Project C, which allows access from Organization A or B. Organization markings, despite being disjunctive within an individual project, propagate along data lineages just like markings, and you need access to all input datasets to view an output dataset, regardless of which project the output dataset resides within.
See this example on creating a shared project across 2 organizations. Note this section related to removing inherited organizations:
After the empty shared Project is set up, developers from both organizations can start referencing data from their private Projects. The organization security primitive protects an organization’s data and propagates along data lineages, similar to Markings. Therefore, any data referenced from one organization won’t be accessible to the other organization. To resolve this, a Sky Industries developer who references a private Sky Industries dataset will have to stop inheriting the upstream Sky Industries organization. The same process applies to datasets private to Sunrise Airline that Sunrise Airline developers reference in the shared Project.