BASH: Truncating file names with special character

tboss888 · 12-18-2008, 06:55 PM

I have a directory with files in it. They have the following naming convention:

file1.txt;1
file2.txt;3
file3.txt;24

I want to remove the ';' and any number that appears after it. Can I use awk or tr to do this? I am thinking of using something like this.

Code:

for x in `ls`
     do
     if [ ! -f $x ]; then
     continue
     fi
     nf=`echo $x | tr <insert code here>
     if [ $nf != $x ]; then
         mv -i $x $nf
     fi
done

Any help appreciated. Thanks.

pixellany · 12-18-2008, 07:14 PM

nf=`sed 's/;[^;]*$//' $x`

matched pattern is "; + any # of chars not ;, at the end of the line". This allows something with more than one ";" (only the last will be removed)

taylor_venable · 12-18-2008, 07:28 PM

I got a killer one-liner:

Code:

find . -name '*;*' -exec bash -c 'mv "{}" "$(echo "{}" | sed -e "s/;[0-9]*\$//g")"' \;

It works for files with spaces in the name (this was the hardest part to figure out), and won't touch files with a ; followed by digits in the middle of the name (only at the end). It's also fully recursive through the directory structure, or only in the current directory or however you wish by tweaking the find flags. Using find is safer than ls because of the variety of conditions under which ls could produce more information than just the filename, for example if ls is aliased to 'ls -l' or such.