Iterate multiple times over all files in a folder and select x % of the files on each run and return the uniquely selected files

Suppose you have a folder with 100 files in it and want to iterate 5 times over these 100 files and return 18 unique files each time in a separate .csv file. What is the most pythonic(python 3.6x) way to do that? the file names saved on each run in a .csv file must be unique(different from each run).