Why is a line item referencing another where both have no overlapping applies to ok?
Lets say I have a line item LineA with an Applies To of Users, and a line item LineB with an Applies To of Organization.
Anaplan lets me make the formula of LineA, "LineB". I understand the principal of using SUM/LOOKUP/SELECT etc. to find the values in each dimension to use where the dimensions are missing from the line item that has the formula, but in the case above I haven't done that (so I'd expect an error). Anaplan lets me do this, so I guess my question is; why? In what cases would the above make sense? Clearly Organization could have many list items each with different values, so how would LineA which doesn't apply to Organization use this?
For some background, I'm creating a Chrome extension to warn about various formula issues, before Anaplan has to try it and do a sometimes costly rollback. For example "formula evaluates to Organization, but the line item type is number". I'm moving on to looking at warning re. miss-matched dimensions and I want to be able to accurately warn about these. To that end I've found a gap in my understanding (the above scenario works, but i'd have thought it shouldn't be allowed).
Re: Why is a line item referencing another where both have no overlapping applies to ok?
Ahh ok, that makes sense thanks.
Yeah, I started working on it a month or so ago (I found the A+ addin and similar didn't do that and seemed fairly limited in that aspect, although helpful). I've since seen that the new editor uses the Monaco editor (used by Visual Studio Code) which is excellent and would enable everything you'd expect; syntax highlighting, code completion, proper live error reporting etc. I had asked if any of that was planned (beyond what's already there re. syntax highlighting) but hadn't got a response yet; is there something I've missed that details further planned features for the formula editor? I had been using the Monaco editor in my extension with a view to add these sorts of things myself after one two many times waiting for the model to roll back an invalid formula. It's quite fun writing the tokenizer / parser etc. so might continue for learning purposes, but it'd be great when/if all the usual code-editing type stuff was included.