Using Linux and Bash, I have a script that outputs filenames with complete path, like this:I would like the following output:And I would like to get the filenames only. Tricky part is that I cannot predict how many levels deep the filename is located, so I cannot use standard Bash-kungfu to solve this...

Add this line to the script

FILENAME=${file##*/}

/foo/bar/filename.pdf
/foo/bar/filename with space.pdf
/foo/filename.txt

Now that worked perfectly!

I have no idea what this means, though. You set a new variable FILENAME that is $file with some magic applied. I do not understand the syntax and I am eager to learn more about this, since I find this very useful. Would you care to explain?

Dare to start out with the man pages ?

Check out the section Parameter Expansion under man sh
and the section Parameters under man ksh

-vino

Found it! I wanted to read more about this, but did not realize this was a shell built-in option, so did not think of man bash (in my case)...

Appreciate the help!

I believe I understand what happened...
So the ##-case expands to the longest matching pattern deletion. Since the parameter is *, it removes all matching patterns in the array, until it has no matches left. You read in the entire pathname as an array which has a delimiter "/" and the final result is the filename only.

Am I right?

This would mean that /foo/bar/filename.pdf would result in filename.pdf, but if the input would have a trailing slash like so: /foo/bar/files/ it would return... nothing!