Are NULLs treated as distinct or duplicates in your UNIQUE INDEX? It depends SQL != NoSQL != some specific implementations Which option offers the best developer experience? Comments welcome