Recursively Counting files by Extension on Mac or Linux

Back in 2004 I wrote up a blog entry showing how to get a count of files by a specific extension. For example you want to know how many js files are in a directory, you can run this:

find /some/dir | fgrep -c '.js'

The -c in the grep command tells it to count the matches. I'm actually using fgrep here because I'm not using regex.

The above would also match a file, or a directory had .js anywhere in the path, so we could improve that script by using a regular expression $ character, for example:

find /some/dir | grep -c '\.js$'

Now we are limiting the .js to show up only at the end of the file. Note, because I'm using grep which uses regex, I now need to escape the . as \.

What if you want a listing of all file extensions and the count of files in a directory?

Here's one way to print out a list of extensions and the number of files of each type:

find /some/dir -type f | grep -o ".[^.]\+$" | sort | uniq -c

This will print out a nice list like this:

5 .js
3 .html
1 .css

How it works

First we have find /some/dir -type f which just limits find to output all the files in the directory recursively. The -type f omits directories from showing up in the list.

Next we have grep -o ".[^.]\+$" the -o tells grep to only output lines that match the pattern, and only output the match. The pattern is just a regex that says look for a dot followed by one or more chars that are not a dot [^.]\+, at the end of a line $.

Next we pipe into the sort command which just puts every thing in order.

Finally we pipe into uniq -c which counts each unique line (the file extensions) and prints out the results. Cool!