i guess, could just be a buggy implementation though. with nostrdb de-dupe is super efficient so its not a big deal. Maybe eventually