How do I count all the files recursively through directories

Question

I want to see how many files are in subdirectories to find out where all the inode usage is on the system. Kind of like I would do this for space usage

du -sh /*

which will give me the space used in the directories off of root, but in this case I want the number of files, not the size.

Solution 1

find . -maxdepth 1 -type d | while read -r dir
do printf "%s:\t" "$dir"; find "$dir" -type f | wc -l; done

Thanks to Gilles and xenoterracide for safety/compatibility fixes.

The first part: find . -maxdepth 1 -type d will return a list of all directories in the current working directory.  (Warning: -maxdepth is a GNU extension and might not be present in non-GNU versions of find.)  This is piped to...

The second part: while read -r dir; do (shown above as while read -r dir(newline)do) begins a while loop – as long as the pipe coming into the while is open (which is until the entire list of directories is sent), the read command will place the next line into the variable dir. Then it continues...

The third part: printf "%s:\t" "$dir" will print the string in $dir (which is holding one of the directory names) followed by a colon and a tab (but not a newline).

The fourth part: find "$dir" -type f makes a list of all the files inside the directory whose name is held in $dir. This list is sent to...

The fifth part: wc -l counts the number of lines that are sent into its standard input.

The final part: done simply ends the while loop.

So we get a list of all the directories in the current directory. For each of those directories, we generate a list of all the files in it so that we can count them all using wc -l. The result will look like:

./dir1: 234
./dir2: 11
./dir3: 2199
...

Solution 2

Try find . -type f | wc -l, it will count of all the files in the current directory as well as all the files in subdirectories. Note that all directories will not be counted as files, only ordinary files do.

Solution 3

Here's a compilation of some useful listing commands (re-hashed based on previous users code):

List folders with file count:

find -maxdepth 1 -type d | sort | while read -r dir; do n=$(find "$dir" -type f | wc -l); printf "%4d : %s\n" $n "$dir"; done

List folders with non-zero file count:

find -maxdepth 1 -type d | sort | while read -r dir; do n=$(find "$dir" -type f | wc -l); if [ $n -gt 0 ]; then printf "%4d : %s\n" $n "$dir"; fi; done

List folders with sub-folder count:

find -maxdepth 1 -type d | sort | while read -r dir; do n=$(find "$dir" -type d | wc -l); let n--; printf "%4d : %s\n" $n "$dir"; done

List folders with non-zero sub-folder count:

find -maxdepth 1 -type d | sort | while read -r dir; do n=$(find "$dir" -type d | wc -l); let n--; if [ $n -gt 0 ]; then printf "%4d : %s\n" $n "$dir"; fi; done

List empty folders:

find -maxdepth 1 -type d | sort | while read -r dir; do n=$(find "$dir" | wc -l); let n--; if [ $n -eq 0 ]; then printf "%4d : %s\n" $n "$dir"; fi; done

List non-empty folders with content count:

find -maxdepth 1 -type d | sort | while read -r dir; do n=$(find "$dir" | wc -l); let n--; if [ $n -gt 0 ]; then printf "%4d : %s\n" $n "$dir"; fi; done

Solution 4

Try:

find /path/to/start/at -type f -print | wc -l

as a starting point, or if you really only want to recurse through the subdirectories of a directory (and skip the files in that top level directory)

find `find /path/to/start/at -mindepth 1 -maxdepth 1 -type d -print` -type f -print | wc -l

Solution 5

du --inodes

I'm not sure why no one (myself included) was aware of:

du --inodes

--inodes
      list inode usage information instead of block usage

I'm pretty sure this solves the OP's problem. I've started using it a lot to find out where all the junk in my huge drives is (and offload it to an older disk).

Further info

If you DON'T want to recurse (which can be useful in other situations), add

-S, --separate-dirs

View more solutions

Comments

[Admin]

See also recursively count all the files in a directory, Count files in each directory? at SU.

[Admin]

I think that "how many files are in subdirectories in there subdirectories" is a confusing construction. If more clearly state what you want, you might get an answer that fits the bill.

[Admin]

@Steven feel free to rewrite it... I thought my example of du -sh /* made it pretty clear how I wanted the count to work. same thing, just count the files not the bytes.

[Admin]

As you mention inode usage, I don't understand whether you want to count the number of files or the number of used inodes. The two are different when hard links are present in the filesystem. Most, if not all, answers give the number of files. Don't use them on an Apple Time Machine backup disk.

[Admin]

@mouviciel this isn't being used on a backup disk, and yes I suppose they might be different, but in the environment I'm in there are very few hardlinks, technically I just need to get a feel for it. figure out where someone is burning out there inode quota.

[Admin]

@mouviciel of course feel free to suggest a way to get the actual inode count

[Admin]

du should provide a builtin feature for this.

[Admin]

@SridharSarnobat, ncdu does have a builtin feature for this, see answer below :)

[Admin]

Actually I have since realized that du provides pretty much a solution built in.

[Johan]

+1 for something | wc -l ... word count is such a nice little tool

[xenoterracide]

yeah but this only does 1 directory.... I'd like to get the count for all directories in a directory, and I don't want to run it seperately each time... of course I suppose I could use a loop... but I'm being lazy.

[Didier Trosset]

find works recursively through all sub directories by default. If you want it to work in multiple locations, you can specify all of them between find and -type.

[xenoterracide]

that second one certainly doesn't work.... I tried it on /home . I got 698035 . I should see about 6 numbers.

[xenoterracide]

way too recursive... I only want to see the top level, where it totals everything underneath it. totaled... this ends up printing every directory.

[Brian Rasmussen]

@xenoterracide: Try adding -maxdepth 1 immediately after the first find. If you want to include the number of subdirectories in your count, remove the -type f at the end (that should have really been ! -type d anyway, so that all non-directory files would have been included).

[Cry Havok]

It works for me - are you sure you only have 6 files under /home? I'd be 100% certain you don't.

[Gilles 'SO- stop being evil']

Always use read -r as plain read treats backslashes specially. Then echo -en "$dir:\t" will again mangle backslashes; a simple fix is to use printf '%s:\t' "$dir" instead. Next, $dir should be "$dir" (always use double quotes around variable substitutions).

[xenoterracide]

@Cry I have maybe 6 directories... not 6 files

[xenoterracide]

modified per @Giles suggestions find -maxdepth 1 -type d | while read -r dir; do printf "%s:\t" "$dir"; find "$dir" | wc -l; done

[Gilles 'SO- stop being evil']

Good idea taking hard links into account. Assuming GNU find, you don't need so many steps: find -printf '%i\n' | sort -u | wc -l. If you wanted to be portable, you'd need find . -exec ls -id {} + | cut … instead.

[xenoterracide]

I'm adding sort -n -r -k2 to the end of this, for lots of directories, so that I know where the most usage is

[DolphinDream]

And btw.. if you want to have the output of any of these list commands sorted by the item count .. pipe the command into a sort : "a-list-command" | sort -n

[Krzysztof Boduch]

The fourth part: find "$dir" makes a list of all the files inside the directory name held in "$dir". You forgot to add -type f to make it list files: find -maxdepth 1 -type d | while read -r dir; do printf "%s:\t" "$dir"; find "$dir" -type f | wc -l; done

[Arsen Karapetjan]

where do I write the name of the root directory?

[MrSpreadsheet]

I'm not getting this to work on macOS Sierra 10.12.5. illegal option -- m in the find command. I changed it to find . -maxdepth ... and got it to work.

[Pablo A]

@xenoterracide you might have dirs with spaces, maybe better setting tab as delimiter sort -rt $'\t' -k 2,2 -V. Also 2>/dev/null after wc pipe to get rid of permission errors.

[x-yuri]

lol, this must be the top-most accepted answer :)

[x-yuri]

Why do you count only files? Directories take inodes as well as files. And well, the directory itself also takes an inode, but that doesn't affect the result much. Anyways, the other answer must be top-most and accepted.

[vstepaniuk]

all those if's could be easily replaced with find [-not] -empty

[sglessard]

On Mac OS, it also counts the .DS_Store inside each directory. You can exclude them with the -type f ! -iname ".*" find parameter. You could also want to exclude files starting by ~*.

[CervEd]

not in freebsd :(

[Sridhar Sarnobat]

I've never used BSD but did you try installing coreutils? I think that gives the GNU version of du. freshports.org/sysutils/coreutils

[Kvothe]

Just to be clear: Does it count files in the subdirectories of the subdirectories etc...?

[herohuyongtao]

@Kvothe Yes, recursively. :)

[netawater]

@DolphinDream thanks, sort -n is good