There aren't many in a production situation. ironically I encountered one just today. It goes like this:
A data model needs to be secured with section access so that individual users can only see the sales transactions that they were responsible for. These transactions have many attributes like location, date, product etc... The section access makes sure that users can only see the possible products, locations etc... that were tied to their transactions...and in this way may only have the ability to toggle between 2 products even if the data model has 100+ products simply because of the section access.
Now the twist is that the dashboard should have global totals at the top. These totals are grand totals for all locations, products etc... and the users can actually filter those based on the dimension values they see. So if they sold 2 products, they can see what the TOTAL sales across all users are for those products.
One way to do this is to have a 2nd data model that is unattached to the first that has many of the same fields as the original data model, but it does not have a user or transaction field in it. The totals are calculated from the 2nd data model and filtered using SET MODIFIERS with values from the first model.
Because there are many like named fields in both the detailed+secured model vs the summary model and you don't want any joins, its easy to just use qualify * for the summary fields to ensure that none of them join .
Thats my example : )
.png "Luminary Alumni"){kind=icon}