0

Some file copying went wrong in one of my machines and I'm wasting a ludicrous amount of space by having a copy of each folder inside of the parent with the same name, and I want to delete them all.

Example:

 /mnt/test/files/foo
 \_ /mnt/test/files/foo/file1 (etc)
 |__ /mnt/test/files/foo/foo
 \_ /mnt/test/files/foo/foo/file1 (etc)
 |_ /mnt/files/foo/foo2
  \_ /mnt/files/foo/foo2/file1 (etc)
 |_ /mnt/files/foo/foo2/foo2
  \_ /mnt/files/foo/foo2/foo2/file1 (etc)

So obviously I want to delete /mnt/files/foo/foo, /mnt/files/foo/foo2/foo2 and its contents (and so on) entirely and stop wasting space. What would be a good way to script it in bash?

SouravGhosh
  • 553
Webodan
  • 3

2 Answers2

2

If your find supports the -regex predicate, you can list the directories with:

find . -type d -regex '.*/\([^/]*\)/\1' -prune -print

to remove them, you can change -print to:

-exec rm -rf {} +

But be sure to check the list first so you don't delete any files you need.

Stéphane Chazelas
  • 544,893
choroba
  • 47,233
0

Using non-GNU find (but still some implementation that supports -mindepth, such as find on BSD systems):

find top-dir -depth -mindepth 1 -type d -exec sh -c '
    for pathname do
        subdir=$pathname/${pathname##*/}
        if [ -d "$subdir" ]; then
            printf "Would remove directory %s\n" "$subdir"
            # rm -rfi "$subdir"
        fi
    done' sh {} +

This would do a depth-first traversal of the directory hierarchy rooted at top-dir. For batches of found directory pathnames, a short shell script would be called. In each iteration of the loop in the short shell script, the pathname for a subdirectory within the $pathname directory that has the same name as the directory itself is constructed. If that subdirectory exists, it is reported (the deletion is currently commented out for safety).

The -depth option causes a depth-first traversal. This is usually what you want when you delete directories with find as find would otherwise potentially try to enter directories that you have already deleted.

The -mindepth 1 option ensures that the top directory is not deleted, as it would be would you use . as the start search path.

Related:

Kusalananda
  • 333,661