How to Count Files in a Directory on Linux and FreeBSD

I was moving backups with rsync. I had no doubts the contents were identical (I’d done it before), but I decided to check the number of files on both sides.

To count the files in a directory and all its subdirectories, use:

ls -l -R /data_b | grep ^- | wc -l

How it works

It’s a compound command, so here it is piece by piece:

• ls -l -R /data_b — prints a detailed list of all files in /data_b and its subdirectories;
• grep ^- — keeps only lines starting with “-“, i.e. regular files (directory lines start with “d”, links with “l”);
• wc -l — counts the lines, i.e. the files.

I’m using FreeBSD as the example because that’s where I needed it, but the commands are standard and work on Linux too.

Note: the total line in the ls -l output is not the number of files but the space they take up (in blocks). People sometimes mistake it for a file count.

Counting by name pattern

You can use regular expressions in grep. For example, the number of .so files in /lib:

ls -l -R /lib | grep ^- | grep '\.so$' | wc -l

More ways: find and ls (Linux)

Later I noted a few more commands for Linux (tested on Debian).

Files in the current directory only, without subdirectories:

find . -maxdepth 1 -type f | wc -l

The fastest option for huge directories is ls -f (no sorting). But it counts every entry: subdirectories, hidden files and the special “.” and “..”.

ls -f . | wc -l

Recursively, including all subdirectories:

find . -type f | wc -l

The number of directories (including the current one) — change f to d:

find . -type d | wc -l

Only files with the .log extension, recursively:

find . -type f -name "*.log" | wc -l