I have a directory with a bunch of files with names like this:
And I want to rename all of them, keeping only the numbers and the extension like this:
How can I do this using a shell script?
I would use (perl-)rename. Distros ship two types of rename, so you can check if you have the perl version of rename installed with
$ rename --version
perl-rename 1.9
Run it with -n to test the command as a "dry run".
$ rename 's/[^0-9]*([0-9]*).*/\1.txt/' * -n
generic name- dd -01 _.txt -> 01.txt
generic name- dd -02 _.txt -> 02.txt
generic name- dd -03 _.txt -> 03.txt
generic name- dd -04 _.txt -> 04.txt
When you are happy run the command itself.
$ rename 's/[^0-9]*([0-9]*).*/\1.txt/' *
rename 's/foo/bar/' *: rename all files (*), replacing the regex foo with bar.[^0-9]*([0-9]*).*: match all non-numerical characters [^0-9]*, followed by numerical characters, which are enclosed in a "capturing group" ([0-9]*), followed by the rest of the characters .*. This will match the whole filename.\1.txt: replace the filename with the captured group \1, appended with .txt.If you have the simpler rename installed, you will see something like this instead.
$ rename --version
rename from util-linux 2.33.1
In this case you can use the following two commands:
$ rename 'generic name- dd -' '' *
$ rename ' _' '' *
But I strongly recommend learning perl-rename instead.
We can do with below one liner. Tested it. But it needs to be executed from the current folder where file is present.
for file in *generic*.txt; do newname=$(echo $file | awk -F"-" '{print $3}' | sed 's/ _//g'); mv "$file" $newname;done
if you want to run from different folder you can try with below script
#!/bin/bash
SRCFLDR=yoursourcefolder
for file in $SRCFLDR/*generic*.txt;do
newname=$(echo $file | awk -F"-" '{print $3}' | sed 's/ _//g')
mv "$file" $newname
echo $file has been renamed to $newname
done
This simple script should do it:
for f in *[0-9]*.*; do
base=${f%.*}
mv -- "$f" "${base//[^0-9]}.${f##*.}"
done
