I suspect the "find all keys with identical values" example would be fastest if they'd use a struct-of-arrays approach where each key gets its own typed array a

I suspect the "find all keys with identical values" example would be fastest if they'd use a struct-of-arrays approach where each key gets its own typed array and objects are identified by a unique index. Just iterate over each array and as soon as a value isn't eqoal to the next then that key is thrown out.