Delete Files Without Corresponding Extension – Shell Script Guide

filesshellwildcards

My goal is to find a Linux command sequence, which deletes some files if a special condition is fulfilled:

All files with the extensions .cut or .cut.bak in the current directory shall be removed, IF there is no file having the same name, but extension .rec or .mpg in the current directory.

Background: I am developing a tool for a Linux-based PVR, that allows cutting of recorded programs. For each recording (extension .rec or .mpg) the segment markers are stored in a .cut-file with the same name. When some recording gets moved/renamed/deleted, then the corresponding cut-file remains orphaned and shall be deleted. I already have implemented the removing of those useless cut-files in C. But I am wondering, whether there may be a (simple) system based solution. In this case it could be run via 'system' and & in the background, which would make my application more responsive…

I already asked a similar question for the recursive case (browsing also into subdirectories), here.
There has been a (nearly) perfect answer by kos – but unfortunately it removes too many files, so it is not usable, and I was not able to figure it out…
But the same idea can be used here, IF it is possible to parse and pipe the result of 'ls' linewise…

Best Answer

Like this?

for f in *.cut *.cut.bak
do
    [ -e "$f" ] || continue
    f="${f%%.bak}"
    f="${f%%.cut}"
    [ -e "$f".mpg ] || [ -e "$f".rec ] || rm -i -- "$f".cut "$f".cut.bak
done

Related Solutions

Unix Command to return all files that end with single digit and TXT extension

One of the most basic tools for locating files (or other kinds of nodes) is the find utility.

find ./ -type f -name '*[!0-9][0-9].txt'

This will search:

...recursivly from the current directory (./). You could change this to another path or even leave it off since this is the default value for most versions of find.
...for items that are files as opposed to directories, devices nodes, symbolic links, etc (-type f). You could leave this off if you wanted to find other types too.
...for items that match the name pattern given. Note that I have used single quotes to encose the pattern so that bash or your shell does not try to expland it to a glob pattern before find even gets the command. The star matches any number of characters, then the end of the file must be something other than a digit, then a digit, then your extention. (-name '*[0-9].txt')

If there are files whose name is only a digit followed by .txt, the command above will miss them, because it requires one non-digit before the digit. The following equivalent commands use boolean operators to include file names with just a digit as well (-o is “or” and ! is “not”):

find ./ -type f \( -name '*[!0-9][0-9].txt' -o -name '[0-9].txt' \)
find ./ -type f -name '*[0-9].txt' ! -name '*[0-9][0-9].txt'

Note that this will be case sensitive. If you would like an insensitive match, you can use -iname instead of -name to match things like file4.TXT as well. Also be aware that just because a file claims to be a text file with that extention does not mean it is. On linux any file can be any type no matter the name. There might also be text files with other extentions or with no extention at all.

Shell – Delete all files starting with a question mark

No need for any fancy stuff. Simply escape the ? so that it's not considered part of the glob:

rm -f ./\?*

This works for ! too:

rm -f ./\!*

Or in one fell swoop:

rm -f ./{\?,\!}*

Update

Just noticed that you were suggesting to grep the output of ls. I wanted to bring your attention to the fact that you shouldn't parse the output of ls

Best Answer

Related Solutions

Unix Command to return all files that end with single digit and TXT extension

Shell – Delete all files starting with a question mark

Related Question