Does this trick preclude the ability to sort your data within a partition? You wouldn’t be able to rely on the row IDs being sequential anymore to be able to just refer to a prefix of them within a newly created file.