mv * folder (avoiding 'cannot move' error) [duplicate]

Question

This question already has an answer here:

• “mkdir foo; svn mv * foo” 3 answers

Anytime I want to move thousands of files to a new folder, I always encounter the same problem.

> mkdir my_folder
> mv * my_folder
mv: cannot move 'my_folder to a subdirectory of itself 'my_folder'

While I think that the error above is harmless (is it?) I am wondering if there is a way of avoiding it.

In case it matters, I am interested in a solution in zsh or one that works well across various shells.

Answer 1

In zsh, with the extended_glob option enabled, you can use ~ to exclude patterns from globs, so you could use:

setopt extended_glob
mv -- *~my_folder my_folder

Or use the negation operator (still with extended_glob):

mv -- ^my_folder my_folder

Use braces to avoid typing the directory name twice:

mv -- {^,}my_folder

In bash (for other answer-seekers using it), you can use Ksh-style extended globs:

# If it's not already enabled
shopt -s extglob
mv -- !(my_folder) my_folder

You can also use that syntax in zsh if you enable the ksh_glob option.

Answer 2

If there are only regular files apart from the new subdirectory, you could use find:

find . -type f -maxdepth 1 -exec mv {} my_folder \;

The -type f option will only find files (not dirs). -maxdepth 1 will keep find looking only in . (not recurs down into other dirs). The -exec … does your move.

In zsh, you can abbreviate this to

mv *(.) my_folder

Yes, the error is harmless.

Answer 3

you can use :

mv * my_folder 2>/dev/null

Answer 4

Actually, there is not problem with that error, the mv command makes its work and you can hide the error redirecting the error output to /dev/null. A simple, but not very elegant workaround is:

mv `ls -A | grep -v my_folder` ./my_folder