Dear Developers,
I would like to propose a new function to determine similarity of sets or collections: the Jaccard similarity index. Such a function is also available in other databases like AsterixDB or ArangoDB. Given the functions intersect(), unionall(), count() and the method .asFloat()such a function can be emulated in ArcadeDB's SQL by something like:

SELECT count(intersect($a, $b)).asFloat() / count(unionall($a, $b)).asFloat() AS jaccard;

given sub-queries $a and $b. However, this is only convenient as an end result, but not as an intermediate step, for example, if based on the similarity something else should happen. Thus, IMHO this function would be a good addition of functionality ArcadeDB currently does not provide.

Thank you for considering this proposition