CODEWITHSUNDEEP
linuxshell
mmo
mmo

Reputation: 4246

Count number of files (with arbitrary filenames) in a given directory

I am trying to count the number of files in a given directory matching a specific name pattern. While this initially sounded like a no-brainer the issue turned out to be more complicated than I ever thought because the filenames can contain spaces and other nasty characters.

So, starting from an initial find -name "${filePattern}" | wc -l I have by now now reached this expression:

find . -maxdepth 1 -regextype posix-egrep -regex "${filePattern}" -print0 | wc -l --files0-from=-

The maxdepth option restricts to the current directory only. The -print0 and the -files0-from options of find and wc, respectively, emit and accept filenames which are null-byte terminated. This is supposed to take care of possible special characters contained in the filenames.

BUT: the --files0-from= option interprets the strings as filenames and wc thus counts the lines contained in those files. But I simply need is the number of files themselves (i.e. the number of null-byte terminated strings emitted by the find). For that wc would need a -l0 (or possibly a -w0) option, which it doesn't seem to have. Any idea how can I count just the number of those names/strings?

And - yes: I realized that the syntax for the filePattern has to be different in the two variants. The former one uses shell syntax while the latter one requires "real" regex-syntax. But that's OK and actually what I want: it allows me to search for multiple file patterns in one go. The question is really just to count null-byte terminated strings.

Upvotes: 1

Views: 40

Answers (1)

John Kugelman
John Kugelman

Reputation: 361605

You could delete all the non-NUL characters with tr, then count the number of characters remaining.

find . -maxdepth 1 -regextype posix-egrep -regex "${filePattern}" -print0 | tr -cd '\0' | wc -c

If you're dealing with a small-to-medium number of files, an alternate solution would be to store the matches in an array and check the array size. (As you touch on in your question, this would use glob syntax rather than regexes.)

files=(*foo* *bar*)
echo "${#files[@]}"

Upvotes: 1

Related Questions