The tarcp.py script uses xz or other compression methods to take the top-level directories and make a 7zip, XZ, or tar archive from each top-level directory, containing all its subdirectories.

Example: say you have 100000 data files scattered by date in subdirectories. This is very cumbersome to share and move around. While I would suggest HDF5 ideally for data archiving, for basic file preservation use my tarcp script above.