Order entries in the CSV dump according to how they are ordered in the genotype files. This way, it's consistent wrt how they appear.