Skip to main content

Hello,
I’ve got a wizard in which the navigation button calls a couple action sequences back to back.

However, I’ve discovered that if I have a validation rule fire or a required field throw an error on the save that is in the first action sequence, the second action sequence is still invoked - and even the navigation at the end of the flow.  So the error flashes, but it bypasses the save, and proceeds with now incorrect data.

Any idea how to get around this?  I’ve played with different approaches but they all move forward in the flow.  Btw, I’ve seen posts on this regarding js, but this is pure skuid wizard/action sequence config.

Any thoughts?

Thanks,
Paul

Hi Paul,

Can you use a branch action to check for a certain field value or other formula that returns true/false, then run the second action sequence only if the formula returns true? Or, you could perhaps have the first action sequence update a UI-only checkbox field if it succeeds, then have the branch action check whether that UI only checkbox is checked or not before running the second action sequence. Let me know if I can clarify this, but I think there should be a way to get the logic you want. 


Hi Mark,

Thanks for the response.


Many of the errors are validation rules, which are used outside of Skuid as well.  As for the branch logic checking a field, it seems really tedious that on every save, I need to set a success value and check for it before moving forward. 


Is there really no better way to error handle in the action framework?  


Thanks!

Paul


Hi Paul, the save action itself offers a way for you to set up on-error actions. However, if you’re stringing multiple action sequences together, that sequence should follow the failure action(s), but it does not pass a failure result to any other action sequence. That’s why I think this particular scenario needs a way to externally track the results of the sequence. Does that help clarify?


Hi Mark.

Seems like a design oversight for the action sequence - asking users to custom build their own error handling logic, state management, and associated testing burden - when a simple “abort” error action option would solve the whole thing, seems very out of character for Skuid.


Is there anyway to convert this from a question to an idea post?


I’ve just converted this to an idea. Thank you for taking the time to explain your scenario. The ability to trigger action sequences with other action sequences (and thus have separate action sequences triggered sequentially) is relatively new, so this is a helpful use case for our product team to know about.