90

I Tried to delete files that starts with A and ends with 2 numbers but It doesn't do a thing.
What I tried:

rm ^A*[0..9]2$

Where am I wrong?

6 Answers 6

104

You can use the following command to delete all files matching your criteria:

ls | grep -P "^A.*[0-9]{2}$" | xargs -d"\n" rm

How it works:

  1. ls lists all files (one by line since the result is piped).

  2. grep -P "^A.*[0-9]{2}$" filters the list of files and leaves only those that match the regular expression ^A.*[0-9]{2}$

    • .* indicates any number of occurrences of ., where . is a wildcard matching any character.

    • [0-9]{2} indicates exactly two occurrences of [0-9], that is, any digit.

  3. xargs -d"\n" rm executes rm line once for every line that is piped to it.

Where am I wrong?

For starters, rm doesn't accept a regular expression as an argument. Besides the wildcard *, every other character is treated literally.

Also, your regular expression is slightly off. For example, * means any occurrences of ... in a regular expression, so A* matches A, AA, etc. and even an empty string.

For more information, visit Regular-Expressions.info.

8
  • 3
    Beware of spaces in file names.
    – slhck
    Feb 22, 2012 at 18:27
  • 1
    The -d"\n switch fixes the spaces problem.
    – Frg
    Feb 22, 2012 at 19:14
  • 3
    Note - some distros (like Mac OS) don't have a grep -P (Perl regex). grep -E may work in this case. Oct 2, 2013 at 20:59
  • 1
    I prefer using -I with xargs and always test with non-lethal commands first: xargs -d"\n" -I {} echo "{}"
    – jozxyqk
    Mar 24, 2014 at 5:40
  • 1
    Parsing ls? See this question which points to this article. Because of the pitfalls you may rm what you don't want to. Nov 15, 2016 at 11:57
81

Or using find:

find your-directory/ -name 'A*[0-9][0-9]' -delete

This solution will deal with weird file names.

6
  • 12
    This is a great solution. I prefer it because it is simpler and you can omit the -delete flag at the end first to see if your regex is correct before mass deleting your files. Apr 14, 2015 at 21:31
  • 2
    Furthermore you have more control on what you delete, for example adding -type f Jun 8, 2016 at 8:12
  • Can this be used to delete files and folders? It does not work for non empty folders.
    – Alex
    Sep 9, 2017 at 11:46
  • @Alex nope, the directory must be empty (it wasn't an OP requirement anyway), you can use the xargs approach with rm -f.
    – cYrus
    Sep 9, 2017 at 13:54
  • also obligatory link to article about parsing ls
    – qwr
    Feb 25, 2020 at 21:48
19

See the filename expansion section of the bash man page:

rm A*[0-9][0-9]
3
  • worked nicely for me. Mar 15, 2017 at 6:09
  • This was the simplest, yet complete answer to the question. Oct 17, 2017 at 17:22
  • Will this work if the folder has a lot of files?
    – Itay
    Sep 3, 2019 at 18:05
4

find command works with regexes as well.

Check which files are gonna to be deleted

find . -regex '^A.*[0-9]{2}$'

Delete files

find . -regex '^A.*[0-9]{2}$' -delete
1
  • This one doesn't work in it's current form. You need to define -regextype egrep to make {x,y} type quantifiers work. Jul 10, 2020 at 15:05
3

This works on my mac:

rm $(ls | grep -e '^A*[0..9]2$')

0
2

The solution with regexp is 200 times better, even with that you can see which file will be deleted before using the command, cutting off the final pipe:

ls | grep -P "^A.*[0-9]{2}$"

Then if it's correct just use:

ls | grep -P "^A.*[0-9]{2}$" | xargs -d "\n" rm

This is 200 times better because if you work with Unix it's important to know how to use grep. It's very powerful if you know how to use it.

2
  • 1
    This doesn't seem to add much beyond what Dennis's 4 year old answer already says.
    – 8bittree
    Nov 16, 2016 at 19:48
  • 1
    "200 times" is a pretty specific. Lots of other commands are very powerful too, all you need to do is learn how to use them. Mar 15, 2017 at 11:22

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .