On Tue, Jun 20, 2023, at 16:56, Joel Jacobson wrote: > I think we have an opportunity here to innovate and potentially influence a > future set concept in the SQL standard.
Adding to my previous note - If there's a worry about future SQL standards introducing SETs with NULLs, causing compatibility issues, we could address it proactively. We could set up set functions to throw errors when passed NULL inputs, rather than being STRICT. This keeps our theoretical alignment now, and offers a smooth transition if standards evolve. Considering we have a flag field in the struct, we could use it to indicate whether a value stored on disk was written with NULL support or not. /Joel
