We know that every finite nilpotent group is written as a direct product of its sylow subgroups. My question is : can we write finite polycyclic groups as a direct product of some subgroups? if the answer was yes then characterize those subgroups.