Joining datasets on a non-distinct attribute
Let's say I need to JOIN 2 datasets, A and B, on an attribute 'key'.
It is distinct for dataset A (each distinct value corresponds to only one example), but not for B (there are few examples with the same value of 'key' in .
What logic of joining will RapidMiner follow in this case? For each join 'A <--> key <--> B' what exactly example will be taken from B? First, last, random, or it totally depends on how dataset was stored in repo so it's not possible to know at all?