Sequentially numbering of files in different folders while keeping the name after the number

Question

I have a lot of ogg or wave files in different folders that I want to sequentially number while keeping everything that stands behind the prefixed number. The input may look like this

Folder1/01 Insbruck.ogg
        02 From Milan to Rome.ogg
        03 From Rome to Naples.ogg


Folder2/01 From Naples to Palermo.ogg
        02 From Palermo to Syracrus.ogg
        03 From Syracrus to Tropea

The output should be:

Folder1/01 Insbruck.ogg
        02 From Milan to Rome.ogg
        03 From Rome to Naples.ogg

Folder2/04 From Naples to Palermo.ogg
        05 From Palermo to Syracrus.ogg
        06 From Syracrus to Tropea.ogg

The sequential numbering across folders can be done with this BASH script that I found here:

find .  | (i=0; while read f; do 
let i+=1; mv "$f" "${f%/*}/$(printf %04d "$i").${f##*.}"; 
done)

But this script removes the title that I would like to keep.

Answer 1

TL;DR

Like this, using find and perl rename:

rename -n 's@/\d+@sprintf "/%0.2d", ++$::c@e' Folder*/*

Drop -n switch if the output looks good.

With -n, you only see the files that will really be renamed, so only 3 files from Folder2.

---

Going further

The variable $::c (or $main::c is a package variable) is a little hack to avoid the use of more complex expressions:

rename -n 's@/\d+@sprintf "/%0.2d", ++our $c@e' Folder*/*

or

rename -n '{ no strict; s@/\d+@sprintf "/%0.2d", ++$c@e; }' Folder*/*

or

rename -n '
    do {
        use 5.012;
        state $c = 0;
        s@/\d+@sprintf "/%0.2d", ++$c@e
    }
' Folder*/*

_{Thanks go|dfish & Grinnz on freenode}

Answer 2

A bash script for this job would be:

#!/bin/bash

argc=$#
width=${#argc}

n=0
for src; do
    base=$(basename "$src")
    dir=$(dirname "$src")
    if ! [[ $base =~ ^[0-9]+\ .*\.(ogg|wav)$ ]]; then
        echo "$src: Unexpected file name. Skipping..." >&2
        continue
    fi
    printf -v dest "$dir/%0${width}d ${base#* }" $((++n))
    echo "moving '$src' to '$dest'"
#   mv -n "$src" "$dest"
done

and could be run as

./renum Folder*/*

assuming the script is saved as renum. It will just print out source and destination file names. To do actual moving, you should drop the # at the beginning of the line # mv -n "$src" "$dest" after making sure it will work as expected. Note that the mv command will not overwrite an existing file due to the -n option. This may or may not be desirable. The script will print out a warning message and skip unexpected file names, that is, the file names not fitting the pattern specified in the question.

Answer 3

The sequential numbering across folders can be done with this BASH script that I found here:

find .  | (i=0; while read f; do 
let i+=1; mv "$f" "${f%/*}/$(printf %04d "$i").${f##*.}"; 
done)

But this script removes the title that I would like to keep.

---

Not as robust as the accepted answer but this is the improved version of your script and just in case rename is not available.

#!/usr/bin/env bash

[[ -n $1 ]] || {
  printf >&2 'Needs a directory as an argument!\n'
  exit 1
}

n=1
directory=("$@")

while IFS= read -r files; do
  if [[ $files =~ ^(.+)?\/([[:digit:]]+[^[:blank:]]+)(.+)$ ]]; then
    printf -v int '%02d' "$((n++))"
    [[ -e "${BASH_REMATCH[1]}/$int${BASH_REMATCH[3]}" ]] && {
      printf '%s is already in sequential order, skipping!\n' "$files"
      continue
    } 
    echo mv -v "$files" "${BASH_REMATCH[1]}/$int${BASH_REMATCH[3]}"
  fi  
done < <(find "${directory[@]}" -type f | sort )

---

Now run the script with the directory in question as the argument.

./myscript Folder*/

or

./myscript Folder1/

or

./myscript Folder2/

or a . the . is the current directory.

./myscript .

and so on...

• Remove the echo if you're satisfied with the output.