Creation of a simple script that renames a file picking randomly from a group of files

Solution 1:

You cannot both "rename" a file and "not lose the original", you'll have to copy (man cp) the randomly selected original file to greetings.ogg, something like:

# select among 0.ogg .. 9.ogg
cp $(( $RANDOM % 10 )).ogg greetings.ogg

Read man bash, you might want to initialize $RANDOM.

Solution 2:

If your files are not necessarily named with a simple arithmetic scheme amenable to the use of the shell's $RANDOM variable, then another option is to use shuf:

shopt -s extglob

cp -- "$(printf '%s\n' !(greeting).ogg | shuf -n 1)" greeting.ogg

The ksh-style extended glob !(greeting).ogg avoids copying the existing file to itself - you could avoid that by copying the file to a different directory.

Solution 3:

A more elegant solution (as already suggested in comments by both Ginnungagap and James S.), would be to use symlinks, like so:

#!/bin/bash

# Select among ten files at random and make a renamed copy inside /home/$user

ln -sf $(( $RANDOM % 10)).mp3 /home/x/greetings.mp3

Note: ln is both faster and more efficient, in terms of disk usage, than cp.

Note the use of the -f option to overwrite any link to a previous greeting sound - see Create symlink - overwrite if one exists.

Alternatively, avoiding the use of $RANDOM (and bash) altogether, and using shuf instead:

#!/bin/sh

rand=`shuf -i 1-10 -n 1`
ln -sf ${rand}.mp3 /home/x/greetings.mp3

The use of shuf is taken from this answer to Generate random numbers in specific range).