Efficient Linux at the Command Line

... character after the asterisk:

$ ls
123  a.txt   b.txt   c.txt  dont-delete-me  important-file  passwords
$ rm * .txt       DANGER!! Don't run this! Deletes the wrong files!

The most common solution to this hazard is to alias rm to run rm -i so it prompts for confirmation before each deletion:

$ alias rm='rm -i'                  Often found in a shell configuration file
$ rm *.txt
/bin/rm: remove regular file 'a.txt'? y
/bin/rm: remove regular file 'b.txt'? y
/bin/rm: remove regular file 'c.txt'? y

As a result, an extra space character needn’t be fatal, because the prompts from rm -i will warn that you’re removing the wrong files:

$ rm * .txt
/bin/rm: remove regular file '123'?      Something is wrong: kill the command

The alias solution is cumbersome, however, because most of the time you might not want or need rm to prompt you. It also doesn’t work if you’re logged into another Linux machine without your aliases. I’ll show you a better way to avoid matching the wrong filenames with a pattern. The technique has two steps and relies on history expansion:

1. Verify. Before running rm, run ls with the desired pattern to see which files match.

   $ ls *.txt
   a.txt   b.txt   c.txt

2. Delete. If the output of ls looks correct, run rm !$ to delete the same files that were matched.²

   $ rm !$
   rm *.txt

The history expansion !$ (“bang dollar”) means “the final word that you typed in the previous command.” Therefore, rm !$ here is shorthand for “delete whatever I just listed with ls,” namely, *.txt. If you accidentally add a space after ...