[SOLVED] How do I remove only sub-directories that themselves only contain a further named sub-directory?

Issue

I have a data directory on a Debian box containing multiple sub-directories, when the system has processed the contents of a sub-directory it adds a sub-directory ".processed" into it.

Unfortunately when the data is deleted the ".processed" directories are left behind and i have to delete them by hand, in among the directories still containing data.

Is there any way to delete only sub-directories that contain just a directory ".processed" and no other files or directories?

Solution

You can use this shell script:

#!/bin/bash
find . -type d -print | tac | while read -r file ; do
    if [[ "$(ls -A "$file")" = ".processed" ]];
    then
      rmdir "$file"/".processed"
      rmdir "$file"
    fi
done

This script loop on all directories given by find . -type d. If it discovers a directory containing only a directory named .processed, then it delete the .processed directory and the directory containing it. Only empty directories are removed, so the script use rmdir. In general, try to avoid using rm -rf in script, because it can be dangerous...

If the order of directories to remove place a subdirectory before a directory then the rmdir will fail. Using rm -rf * solves the issue but is dangerous. The solution here is to pipe to tac and reverse the order given by find.

The script works with white spaces in file names. Thanks to the solution from the Sorpigal answer, named pipe to while, newline terminated.

Answered By - Ortomala Lokni