[SOLVED] Small scripting query: lowercase

Steve W · 03-23-2011, 04:36 AM

I have several directories of subtitles for some videos, but all the text for the subtitles in each file is in uppercase, and I would like to convert the files' content to lowercase, all in one go. I found on a website a bash command that would do each file separately:

tr '[:upper:]' '[:lower:]' <input.txt> output.txt

... but of course involves specifying the input name and output name for each file.

I have been trying to do it in a script that would work for all files in the current directory, without having to rename them each time, if that's possible. So far I've got the following, which doesn't work:

#!/bin/bash

for file in $@; do
tr '[:upper:]' '[:lower:]'
done

However, my web searches cannot locate a way of specifying "each file in current directory" in conjunction with the "tr" command, but also without having to rename the file once converted to lowercase. Is it possible, or would the tr command have to create a new file for each converted file?

Thank you for any advice you can give.

chrism01 · 03-23-2011, 05:22 AM

Code:

for file in *.txt
do
    tr '[:upper:]' '[:lower:]' <$file >${file}.new
    mv $file.new $file
done

tr is a stream oriented editor; it can't replace file contents in-place.
You can prob use sed -i ... if you're that bothered about creating tempfiles.

grail · 03-23-2011, 05:24 AM

Well the first thing to ask is do any of the file names contain spaces? If yes then you will need a while loop as opposed to a for loop which will break the words on whitespace.

As for what to use in the loop, I am sure there are many methods, but the following would work:

Code:

sed -i 's/.*/\L&/' file

Give it a go on a single file to see if you get the desired result first. Should you want a backup of the original file, simply add backup extension to -i, eg -i.bak

Steve W · 03-23-2011, 11:53 AM

Thank you both for the advice. Files lowercased!