Running a bash script recursively and performing operations on all files within the subdirectories

Question

I'm trying to convert flac files into wav files using ffmpeg. The flac files are located in various subdirectories.

/speech_files
/speech_files/201/speech1.flac
/speech_files/201/speech2.flac
/speech_files/44/speech45.flac
/speech_files/44/speech109.flac
/speech_files/66/speech200.flac
/speech_files/66/speech33.flac

What I want after the script runs is the following

/speech_files
/speech_files/201/speech1.wav
/speech_files/201/speech2.wav
/speech_files/44/speech45.wav
/speech_files/44/speech109.wav
/speech_files/66/speech200.wav
/speech_files/66/speech33.wav

I can get my script to work within one directory but I'm having a hard time getting it to run from the top level directory (speech_files) and work it's way through all the subdirectories. Below is the script I'm using.

#!/bin/bash

for f in "./"/*
do
    filename=$(basename $f)
    if [[ ($filename == *.flac) ]]; then
        new_file=${filename%?????}
        file_ext="_mono_16000.wav"
        wav_file_ext=".wav"
        ffmpeg -i $filename $new_shits$wav_file_ext
        ffmpeg -i $new_file$wav_file_ext -ac 1 -ar 16000 $new_file$file_ext
        rm -f $filename
        rm -f $new_file$wav_file_ext
    fi
done

When you want to `cd` to a dir with flac's, look at https://stackoverflow.com/a/42683508/3220113 — Walter A, Jun 17 '18 at 15:44

Idriss Neumann · Answer 1 · 2018-06-17T15:08:12.913

0

Using bash only :

#!/bin/bash

DIR="/.../speech_files"

process() {
    filename=$(basename "$1")
    # ...
}

for f in n "${DIR}"/*/*.flac; do
    process "$f"
done

Using find which is recursive and more efficient to do that kind of task to me :

find "${DIR}" -type f -a -iname "*.flac" -exec ... {} \;

edited Jun 17 '18 at 15:08

answered Jun 17 '18 at 15:02

Idriss Neumann

3,760
2
23
32

score 0 · Answer 2 · answered Jun 17 '18 at 15:02

0

Use find from the top level directory and filter by using *.flac.

for f in $(find . -name "*.flac"); do
    echo "$f" # f points to each file
    # do your logic here
done

answered Jun 17 '18 at 15:02

apatniv

1,771
10
13

No need to use a `for` loop to iterate on a `find` output, you have the `-exec` option which is a lot safer. `for i in $(command)` might cause bugs when the filename contains some spaces for example. – Idriss Neumann Jun 17 '18 at 15:04
Yes, there are better options to make something faster. If writing faster code is the only criteria, then I would ask him/her in something like "Java" or c++ and do all the hard work. – apatniv Jun 17 '18 at 15:06
It's not only a matter of doing things faster, `for i in $(command)` is a bad practice that often cause bugs (for example it's not working when a filename contain spaces). It's also a matter of not re-inventing the wheel: if you choose to use `find` the use `find` completely. – Idriss Neumann Jun 17 '18 at 15:12
If you really need to iterate on an output of command (it's not the case here), there are many safer ways and I advice you to read this : https://stackoverflow.com/a/19607361/2900196 – Idriss Neumann Jun 17 '18 at 15:16
1

I found a nice answer here: [How to loop through file names returned by find?](https://stackoverflow.com/a/9612232/6176817) – PesaThe Jun 17 '18 at 15:19
@PesaThe Indeed :) – Idriss Neumann Jun 17 '18 at 15:24

Running a bash script recursively and performing operations on all files within the subdirectories

2 Answers2