Also hashed based partitioning is incorrect. If you were to use ID then you could end up lopsided if the IDs are heavily weighted to one partition unless IDs are randomly generated. Instead you should always hash the ID so that it's evenly distributed.