I have a directory full of images:
image0001.png
image0002.png
image0003.png
...
And I would like a one-liner to rename them to (say).
0001.png
0002.png
0003.png
...
How do I do this?
On Debian and derivatives, Perl's rename commandline works similarly to sed like this:
rename -v 's/image//' ./*.png
There's also the rename from util-linux that works like this, instead:
rename -- image '' *.png
rename program came from a Perl examples distribution. Debian and Ubuntu ship it as /usr/bin/rename. Other Unix variants may not provide it, or may provide a completely different /usr/bin/rename.
rename take a from to replacement pattern: rename [options] <expression> <replacement> <file>...
If you are using Bash or other POSIX-compatible shell:
for f in *.png; do
mv -- "$f" "${f#image}"
done
The zsh shell has a powerful batch rename command called zmv.
First you need to enable the zmv command as follows (this can go into your ~/.zshrc).
autoload zmv
The basic syntax is zmv PATTERN REPLACEMENT. The pattern is a shell glob expression. Parts of the pattern can be surrounded by parentheses. The replacement text can contain $1, $2, etc. to refer to the Nth parenthesised group in the pattern. For example:
zmv 'image(*.png)' '$1'
You can also ask zsh to automatically define $1, $2, etc. to match the wildcard characters in the pattern:
zmv -w 'image*.png' '$1.png'
zmv prezto requires uncommenting the line in your ~/.zpreztorc following # Set the Zsh functions to load (man zshcontrib). giving zstyle ':prezto:load' zfunction 'zargs' 'zmv'
autoload zmv in your ~/.zshrc: you can also just run it whenever you need zmv.
I normally use the nice and simple mmv (man page) utility for this usecase:
$ mmv "image*.png" "#1.png"
will perform your task.
The #1 in the target pattern will be substituted with whatever matches the wildcard in the source pattern. This also works for several wildcards and can be used for example to change the order of parts of filenames. You can also easily do more complicated things like converting lower case to upper case letters.
Make sure to protect the patterns from the shell by quoting.
The command qmv from renameutils opens an editor showing a list of filenames with two colums, separated by a tab. Each row shows one of the filenames, the same in both columns. The right column is representing the new names of the files.
To make changes, edit the names on the right side. In this example, :%s/... or visual block mode are helpful.
$ qmv *.png
In editor:
image0001.png image0001.png
image0002.png image0002.png
image0003.png image0003.png
~
~
~
~
"/tmp/user/1000/qmvxWyVMs" 3L, 93C
Edit names in right column:
(Removing the image prefix from all lines using visual block mode)
image0001.png 0001.png
image0002.png 0002.png
image0003.png 0003.png
~
~
~
~
:wq
Log of renaming:
image0001.png -> 0001.png
image0002.png -> 0002.png
image0003.png -> 0003.png
(e.g. Ubuntu: apt-get install renameutils)
Uses sed to rename
for i in image*.png
do
mv -i -- "$i" "$(printf '%s\n' "$i" | sed '1s/^image//')"
done
I like Perl so:
perl -nlE '$old=$_; s/image//; qx(mv $old $_)'
You can also use the same pattern for other tasks like copying the files to another directory:
perl -nlE '$old=$_; s(image)(/path/to/new/dir/); qx(mv $old $_)'
rename($old,$_)
easy recurse selecting image*png files, and assumes no need to deal with newline, backslash in file names
find . -name "image*.png" | while read f; do mv -v "$f" "$(echo "$f" | sed -e 's/image//' - )"; done
echo not modifying data, which it could do if it contains backslashes under some circumstances.
Using a shells that supports brace expansion (Bash, Zsh, Ksh93, Fish, Csh, Tcsh):
for N in {0001..1000}; do mv "{image,}$N.png"; done
Strangely, no one mentioned this well covered approach:
find . | grep \.png$ | sed 'p;s/image//' | xargs -n2 mv
And if you like playing around with arguments:
find . | grep \.png$ | sed "p;s/image//" | xargs -n2 sh -c 'echo $1 $2' $0
Single quotes after sh -c matter
$. Should be grep '\.png$'. (3) s/image// removes the first image in the path, so for a ./my-images/image0001.png that would try and do mv ./my-images/image0001.png ./my-s/image0001.png and fail. You'd want to process the list depth first and only do the rename on the basename.
$0, $1,$2.
\. doesn’t do what you think it does: echo \. prints ., without the backslash. Also, why waste resources piping find into grep when you can use -name?
Try brename (https://github.com/shenwei356/brename), a practical cross-platform command-line tool for safely batch renaming files/directories via regular expression (supporting Windows, Linux and OS X) .
@patrickDurusau said:
Linux has a variety of batch file renaming options but I didn’t see any short-comings in brename that jumped out at me.
Features:
find ./ -name "*.html" -exec CMD.Command:
$ brename -f .png -p image
[INFO] checking: [ ok ] 'image0001.png' -> '0001.png'
[INFO] checking: [ ok ] 'image0002.png' -> '0002.png'
[INFO] checking: [ ok ] 'image0003.png' -> '0003.png'
[INFO] 3 path(s) to be renamed
[INFO] renamed: 'image0001.png' -> '0001.png'
[INFO] renamed: 'image0002.png' -> '0002.png'
[INFO] renamed: 'image0003.png' -> '0003.png'
[INFO] 3 path(s) renamed
For Windows and linux, this Perl script will do; in this case:
$ rnm -l 's/^image//' '*.png'
The script could run recursively under directories and even prepending a count to all of them:
$ rnm -r 's/^/$counter./' '/\.png$/'
UTF-8 chars are also correctly treated, both in Windows and linux.
rnm was already suggested in 2015 in a separate answer.
Reading names from find command.
find . ! -path . -prune -type f -name 'image*png' |
while IFS= read -r f; do
mv "$f" "$(printf '%s\n' "$f" | sed -e 's/^\.\/image//' - )"
done
Reading names from a file
while IFS= read -r f; do
mv "$f" "$(printf '%s\n' "$f" | sed -e 's/^\.\/image//' - )"
done < flist
Both of these approaches assume pathnames have no embedded newlines.
thunar -B *.pngfor thunar's bulk-rename gui tool.