Can someone tell me if the two data models will ever yield different results?
What I've done is created test data with ctrl-00 then duplicated the data but changed the name of the keys.
In the first instance all the tables are joined by the "num" column. In the second instance I've duplicated the "num" column and renamed it to "num_ascii2" and "num_characters"
They can give different results if a num value exists in both the Characters and Ascii tables, but not in the Transactions table. In the left data model Ascii and Characters are still associated. In the data model on the right they're not. So you wouldn't be able to see the related data in the Ascii table if you select a num value in Characters that doesn't exist in Transactions, but does in Ascii
That makes sense. Thanks. So the preferred method would be the data model on the left where all tables are associated?
That depends. Perhaps you have a good reason to prevent a direct association between the Ascii and the Characters table.