I have this huge folder with thousands of unordered files. Is it feasible to move the first 5000s to a subfolder via the mv command? For now I move files with
mv *some_pattern* ./subfolder1/ As for now, I move images quite randomly, it's not really important if there aren't exactly 5000 files in each subfolder. Is there a better way to do it?
mv `ls | head -500` ./subfolder1/ 
- or . and assuming subfolder1 itself does not show up in that list.) $ sudo mv `ls | head -50000` 01/ sudo: unable to execute /bin/mv: Argument list too long . Something like: mv ls | head -500` /subfolder1/` $IFS, as my files contained spaces. <insert eye-roll here> With zsh and its glob qualifiers:
mv -- *(D.oN[1,5000]) ./subfolder1 To move up to 5000 regular files in the order they are in the directory.
For the first 5000 in the lexicographically sorted list:
mv -- *some_pattern*(D.[1,5000]) ./subfolder1 If you get an error about arg list too long. You can use zsh's buitin mv command by issuing:
zmodload zsh/files first.
POSIXly:
set -- for f in .*some_pattern* *some_pattern*; do [ "$#" -lt 5000 ] || break [ -f "$f" ] || continue [ -L "$f" ] && continue set -- "$@" "$f" done [ "$#" -eq 0 ] || mv -- "$@" subfolder1/ On a GNU system, you could also do:
find . ! -name . -prune -name '*some_pattern*' -type f -print0 | head -zn 5000 | xargs -r0 mv -t ./subfolder1 -- 
-- *(D.oN part? Thanks! A version that is simple and supports special chars, spaces, etc.
ls -Q dir1 | head -1000 | xargs -i mv dir1/{} dir2/ For this to work as-is dir2 must exist and you have to execute it from the parent directory of dir1 and dir2.
This will move 1000 files from dir1 to dir2.
ls -Q -S dir1 | head -1000 | xargs -i mv dir1/{} dir2/ for moving 1000 largest files in dir1 (-S lists file by size) ls -Q does not produce an output compatible with xargs's expected input format. It helps for file names containing space characters, but not for double quotes or backslashes and harms for file names containing control characters including TAB. find . -name 'Hello*.gz' | head -n 5000 | xargs -d $'\n' mv -t /data01/path/ In the find command, . (dot) denotes current directory
Finds files which start with Hello and end with .gz, first 5,000 files will be moved to /data01/path/.
You might need to do something like this:
x=1 for file in * do if [ "X$x" = "X#####" ]; then break fi mv $file <destination> x=`expr $x + 1` done This script works in bash, ksh, sh and multiple UNIX variants.
destination itself does not show up in that list.) I was able to succesfully do a move of 50,000 without the mv bash error like this
ls | head -50000 | xargs -I{} sudo mv {} 01/ Funny enough, this was on a samba share, so the 50k limit is because Windows Explorer GUI does not like more than 60k files in a dir in general.
zsh: argument list too long: mv in my case ls i | head -500000 | xargs -I {} mv i/{} o/{} Version working even in extremely large directories: moving thousand files from directory containing half million:
ls -U . | head -1000 | xargs mv -t 01/ (Note that the syntax with {} is spawning new mv for every file, while this one doesn't)
