Matching filename in ls (bash)

Question

I have the following files

tcpdump-12
tcpdump-12.delay
tcpdump-24
tcpdump-24.delay

Is there a way to ls only the files

tcpdump-12
tcpdump-24

I can do

ls tcpdump-[[:digit:]][[:digit:]]

but I am looking for something more generic that can take any number of digits, something like tcpdump-[0-9]+ if I was using vim or python regular expressions.

Answer 1

One needs to turn on extended glob functionality of bash to be able to use the advanced pattern matching.

$ ls
tcpdump-12  tcpdump-12.delay  tcpdump-24  tcpdump-24.delay
$ shopt -s extglob
$ ls tcpdump-+([[:digit:]])
tcpdump-12  tcpdump-24

Answer 2

如果您确定所有不需要的文件以“.delay”结尾，则可以执行以下操作：

 ls --ignore '*.delay' 

Answer 3

I'm not sure why you're using [[:digit:]] rather than [0-9] ; are you concerted the file names might contain other kinds of digits?

Most of the other answers are good, but a quick-and-dirty solution is:

ls tcpdump-*[0-9]

It works for the particular set of files you have, but it would also match file names like tcpdump-FOO7 .

In a general-purpose script, it's worth the effort to match exactly the pattern you want. In a one-short interactive shell command, sloppy shortcuts that just happen to work for the current situation can be useful.

Answer 4

You could pipe the output from ls into grep. Grep has an "invert" option (to show lines that don't match), so you could do this:

 ls tcpdump-* | grep -v '\.delay$'

Answer 5

If you don't mind passing them through an external filter, you can use:

ls -1 tcpdump-* | grep '^tcpdump-[0-9]*$'

Just keep in mind this ends up giving you one per line rather than a nice multi-columnar ls output.

If you're processing that list, it wont matter too much but, if you just want to see those files in a directory listing, it's not good enough. I'm assuming the former since, otherwise, this question doesn't really belong on SO :-)