This is a spin-off out of another problem here for a proposal of improvement.

It could actually be beneficial for built-in accumulate functions to ignore NULLs.

This would be similar behavior of other languages such:

• SQL (e.g.: MySQL, Oracle, ..)
• Clojure

Currently it is up to the rule author to put a guarding condition in the first accumulate <source pattern> to manually ignore null sources to avoid the built-in accumulate functions to break trying to accumulate on NULLs.