I bump into this now and then and thought I'd probe the list for an explantion that will stick in my memory.
Imagine two tables where three fields are used in a multi-predicate join to generate a dynamic value-list.
FieldC in Table2 sometimes has data and sometimes not. When Table1FieldC is populated the join works when Table2FieldC also has data. But the whole join breaks when both are empty. (FieldA and FieldB always have data)
I keep thinking that nothing=nothing or empty=empty.
I can "trick" the join by creating FieldD that is an auto-enter and replace contents when either FieldB or FieldC change and then making the join as:
An even chintzier cheat is to set FieldC (both tables) to auto-enter a space character and then use a field menu control in both tables (1,2,3) to set data into it.
I gather that one of the rules is that data must exist in a field for it to validate a relationship. I just keep getting surprised when having an "empty" predicate breaks the whole thing.
Wondering if there is an appropriate techno-term or rule that describes that condition. "Fudd's First Law of Opposition"? or "Wonko's Law of TV Rerun Likelihood"? Perhaps that will help me remember...
"The time from now until the completion of the project tends to become constant." --- Douglas Hartree