You will encounter them from time to time: files with weird filenames, such as spaces, escape codes, or uncommon characters. These often can be very difficult to remove.

For example, files with a space at the end:

# touch "a file "
# ls
a file

It's not such a problem, if you created the file yourself, and you KNOW there is space at the end. Otherwise, it can be quite difficult to remove it:

# rm "a file"
rm: a file: A file or directory in the path name does not exist.

It can be even more ugly if there is a ^M in the file name:

# touch 'a^Mfile'
# ls a*
a
file
# ls file
ls: 0653-341 The file file does not exist.

And it will quickly become horrible if there are unprintable characters in file names, or a combination of all of the above. Or how about a file called "-rf /". Would do dare run the command: "rm -rf /" on your system, not knowing if this will wipe out all files, or just remove the file with the filename "-rf /"?

So, if you have a file with an awkward filename, or simply don't know the file name of a file, because it contains unprintable characters, escape codes, slashes, spaces or tabs, how do you safely remove it?

Well, you can remove files by inode. First, discover the inode of a file:

# ls -alsi
12294  0 -rw-r--r--  1 root  system  0 May 07 15:38 a file

In the example above, the inode has number 12294. Then simply remove it using the find command, and the -exec option for the find command:

# find . -inum 12294 -ls
12294  0 -rw-r--r--  1 root  system  0 May  7 15:38 ./a file
# find . -inum 12294 -exec rm {} \;

If you found this useful, here's more on the same topic(s) in our blog:

• TCPdump/IPtrace existing files?
• Temperature monitoring
• Prune old Docker data
• Creating a CSV file from NMON data
• How to migrate from p5 to p6