How to rename to fixed length

Question

I have files named as 0-n.jpg for n from 1 to 500, for example.

The problem is that some guy using Windows didn't use leading zeros so when I do ls I obtain

0-100.jpg
0-101.jpg
...
0-10.jpg
...
0-199.jpg
0-19.jpg
0-1.jpg

So I'd like to rename them to insert the leading zeros, so that the result of ls could be

0-001.jpg
0-002.jpg
...
0-100.jpg
...
0-499.jpg
0-500.jpg

In other words, I'd like all files with the same file name lengths.

I tried this solution but I'm getting a sequence of errors like

bash: printf: 0-99: invalid number

Answer 1

If your system has the perl-based rename command you could do something like

rename -- 's/(\d+)-(\d+)/sprintf("%d-%03d",$1,$2)/e' *.jpg

Testing it using the -v (verbose) and -n (no-op) options:

$ rename -vn -- 's/(\d+)-(\d+)/sprintf("%d-%03d",$1,$2)/e' *.jpg
0-10.jpg renamed as 0-010.jpg
0-19.jpg renamed as 0-019.jpg
0-1.jpg renamed as 0-001.jpg

Answer 2

If you have zsh shell, you could do something like below.

zmv '([0-9])-([0-9]##).(jpg)' '$1-${(l:3::0:)2}.$3'

Testing

touch 0-1.jpg
touch 0-23.jpg
touch 0-345.jpg
touch 0-6.jpg
touch 0-05.jpg

Change the shell to zsh and if zmv is not loaded, you could do autoload zmv.

Now, you could add -n flag to the zmv command to see what will happen if you execute the zmv command. I am running the command as,

zmv -n '([0-9])-([0-9]##).(jpg)' '$1-${(l:3::0:)2}.$3'

The output is,

zmv -n '([0-9])-([0-9]##).(jpg)' '$1-${(l:3::0:)2}.$3'
mv -- 0-05.jpg 0-005.jpg
mv -- 0-1.jpg 0-001.jpg
mv -- 0-23.jpg 0-023.jpg
mv -- 0-6.jpg 0-006.jpg

If you are satisfied the file names are renamed correctly, you could remove the -n flag.

References

Thanks to user Gnouc for clarifying my doubt on this question which I had posted as another question to get this solution and the link to Gnouc's answer is this one.

Answer 3

another solution in case you do not have "rename:

for file in *.jpg; do
   [ -f "$file" ] || continue  #skips if no jpg file present in current dir
   echo "${file}" | awk -F'[-.]' '
      {   new=sprintf("%03d",$2); 
          print "echo TESTING mv ", $0," ",$1"-"new"."$3 
      }'
done | bash

(edit: moved the | bash so it's only invoked once, and not once per file. And this quick loop is not "weird filename" friendly, but could be sanitized further (to handle files with spaces, etc))

and remove the 2 words echo TESTING to really do the mv ..., once you are sure it's doing what you want it to do.

Answer 4

Here's another approach. Since you only need to rename a specific subset of files and since you already know what those are, just list them and rename them:

for i in {1..99}; do mv -- 0-$i.jpg 0-$(printf '%03d\n' $i).jpg; done

That will list the numbers from 1 to 99, saving each value as $i. You don't care about the rest since those are already named correctly. You then rename 0-$i.jpg to the padded version. The printf %03d directive will print a number adding 0s to the left hand side if it is less than 3 characters in length. The result is that all your problematic files will be renamed.