1456

I need to get a list of human readable du output.

However, du does not have a "sort by size" option, and piping to sort doesn't work with the human readable flag.

For example, running:

du | sort -n -r 

Outputs a sorted disk usage by size (descending):

du |sort -n -r
65108   .
61508   ./dir3
2056    ./dir4
1032    ./dir1
508     ./dir2

However, running it with the human readable flag, does not sort properly:

du -h | sort -n -r

508K    ./dir2
64M     .
61M     ./dir3
2.1M    ./dir4
1.1M    ./dir1

Does anyone know of a way to sort du -h by size?

6

39 Answers 39

2059

As of GNU coreutils 7.5 released in August 2009, sort allows a -h parameter, which allows numeric suffixes of the kind produced by du -h:

du -hs * | sort -h

If you are using a sort that does not support -h, you can install GNU Coreutils. E.g. on an older Mac OS X:

brew install coreutils
du -hs * | gsort -h

From sort manual:

-h, --human-numeric-sort compare human readable numbers (e.g., 2K 1G)

28
  • 4
    The relevant section of the manual: gnu.org/software/coreutils/manual/…
    – wodow
    Feb 9, 2011 at 11:13
  • 32
    Easy to install on OS X with homebrew -- brew install coreutils. May 1, 2011 at 18:53
  • 50
    Good one! I personally always did du -BM | sort -nr as a workaround - it is human readable enough, and it is sorted, if anyone is stuck with older coreutils.
    – chutz
    May 24, 2012 at 9:06
  • 32
    If using on OSX via Homebrew, note that you'll now need to use gsort rather than sort: du -hs * | gsort -h Dec 16, 2013 at 8:45
  • 2
    @PaulDraper, du -BM prints everything in megabytes, so a file that is 168K would actually display as 0M. Unless there is some other version discrepancy I am not aware of. My version of du only shows integer megabyte values.
    – chutz
    Feb 1, 2014 at 15:12
111
du | sort -nr | cut -f2- | xargs du -hs
8
  • 59
    And it'll do a huge amount of duplicate counting. Feb 25, 2009 at 13:55
  • 1
    First it does the normal du - then for each entry it recalculates the size just to print it in human readable form. Feb 25, 2009 at 14:22
  • 8
    @Douglas Leeder: you are right for the duplicate counting, but think that the second du does not start from cold cache (thanks to the OS) @hasen j: xargs is a very useful command, it splits its stdin and feeds it as arguments to the given command
    – cadrian
    Feb 25, 2009 at 14:52
  • 4
    Chris's is actually superior since it works with paths containing whitespace. Throwing a vote your way, buddy.
    – rbright
    Feb 25, 2009 at 22:45
  • 3
    Ugly, but cross-platform :).
    – voretaq7
    Nov 29, 2011 at 23:06
86

There is an immensely useful tool I use called ncdu that is designed for finding those pesky high disk-usage folders and files, and removing them. It's console based, fast and light, and has packages on all the major distributions.

3
  • Very nice... I wondier if the results could be fed to standard out... I am so lazy that I cannot read the manual
    – ojblass
    Jun 27, 2009 at 5:17
  • 8
    gt5 is in the same vein; its killer feature is displaying growth.
    – Tobu
    Jul 1, 2010 at 16:16
  • 2
    That's really cool! And much faster than hanging around with du, if you just want to identify the large directories.
    – BurninLeo
    Sep 16, 2015 at 17:49
70

@Douglas Leeder, one more answer: Sort the human-readable output from du -h using another tool. Like Perl!

du -h | perl -e 'sub h{%h=(K=>10,M=>20,G=>30);($n,$u)=shift=~/([0-9.]+)(\D)/;
return $n*2**$h{$u}}print sort{h($b)<=>h($a)}<>;'

Split onto two lines to fit the display. You can use it this way or make it a one-liner, it'll work either way.

Output:

4.5M    .
3.7M    ./colors
372K    ./plugin
128K    ./autoload
100K    ./doc
100K    ./syntax

EDIT: After a few rounds of golf over at PerlMonks, the final result is the following:

perl -e'%h=map{/.\s/;99**(ord$&&7)-$`,$_}`du -h`;die@h{sort%h}'
9
  • 2
    Your short version outputs on stderr because of the die can you change it to make it output on stdout? Sep 4, 2009 at 16:16
  • 2
    Change the die to a print and it will go to stdout. It's just two more characters. Sep 9, 2009 at 17:55
  • works on ubuntu!
    – marinara
    Apr 26, 2012 at 8:22
  • impressive perl hackistry
    – nandoP
    Nov 13, 2013 at 5:00
  • The result is in reverse order :(
    – RSFalcon7
    Apr 1, 2014 at 22:28
48
du -k * | sort -nr | cut -f2 | xargs -d '\n' du -sh
3
  • Can't use with du -k --total, gives error at the end du: cannot access 'total': No such file or directory Jan 3, 2015 at 2:51
  • i like this one more any other answer. how would you go to show only the first 50 results?
    – Mau
    Jan 12, 2016 at 4:25
  • 1
    @Mauro -- just pipe the result to head by adding ` | head -50` at the end. Feb 21, 2018 at 10:32
24

As far as I can see you have three options:

  1. Alter du to sort before display.
  2. Alter sort to support human sizes for numerical sort.
  3. Post process the output from sort to change the basic output to human readable.

You could also do du -k and live with sizes in KiB.

For option 3 you could use the following script:

#!/usr/bin/env python

import sys
import re

sizeRe = re.compile(r"^(\d+)(.*)$")

for line in sys.stdin.readlines():
    mo = sizeRe.match(line)
    if mo:
        size = int(mo.group(1))
        if size < 1024:
            size = str(size)+"K"
        elif size < 1024 ** 2:
            size = str(size/1024)+"M"
        else:
            size = str(size/(1024 ** 2))+"G"

        print "%s%s"%(size,mo.group(2))
    else:
        print line
0
24

I've had that problem as well and I'm currently using a workaround:

du -scBM | sort -n

This will not produce scaled values, but always produce the size in megabytes. That's less then perfect, but for me it's better than nothing (or displaying the size in bytes).

2
  • I like th -BM switch, which is basically the same as -m, but it has the advantage of displaying the size and M postfixed to it, so you get 10M which is much clearer than just 10 :)
    – Tom Feiner
    Feb 25, 2009 at 14:02
  • This is the simplest solution I've seen so far on this page, thank you!
    – Jeff Olson
    Nov 5, 2015 at 16:35
21

Found this posting elsewhere. Therefore, this shell script will do what you want without calling du on everything twice. It uses awk to convert the raw bytes to a human-readable format. Of course, the formatting is slightly different (everything is printed to one decimal place precision).

#/bin/bash
du -B1 | sort -nr  |awk '{sum=$1;
hum[1024**3]="G";hum[1024**2]="M";hum[1024]="K";
for (x=1024**3; x>=1024; x/=1024){
        if (sum>=x) { printf "%.1f%s\t\t",sum/x,hum[x];print $2;break
}}}'

Running this in my .vim directory yields:

4.4M            .
3.6M            ./colors
372.0K          ./plugin
128.0K          ./autoload
100.0K          ./syntax
100.0K          ./doc

(I hope 3.6M of color schemes isn't excessive.)

5
  • 3
    I have a Perl answer, too, but I think it might make people hate me: du -B1 | sort -nr | perl -e '%h=(0=>b,1=>K,2=>M,3=>G);for(<>){($s,@f)=split/\s+/;$e=3;$e-- while(1024**$e>$s);$v=($s/(1024**$e));printf "%-8s%s\n",sprintf($v >= 100 ? "%d%s" : "%.1f%s",$s/(1024**$e),$h{$e}),@f;}' Feb 25, 2009 at 14:40
  • Even though the Perl answer actually gives its formatting much closer to du. Although the rounding is off... It looks like du always gives ceil() rather than round() Feb 25, 2009 at 14:41
  • Hey, why did I use a hash there? Should've been an array... morning-brain grumble.... Feb 25, 2009 at 15:33
  • Added a better Perl solution as another answer. Feb 25, 2009 at 21:06
  • Both versions fail when filenames contain spaces
    – Vi.
    Nov 2, 2010 at 17:54
20

This version uses awk to create extra columns for sort keys. It only calls du once. The output should look exactly like du.

I've split it into multiple lines, but it can be recombined into a one-liner.

du -h |
  awk '{printf "%s %08.2f\t%s\n", 
    index("KMG", substr($1, length($1))),
    substr($1, 0, length($1)-1), $0}' |
  sort -r | cut -f2,3

Explanation:

  • BEGIN - create a string to index to substitute 1, 2, 3 for K, M, G for grouping by units, if there's no unit (the size is less than 1K), then there's no match and a zero is returned (perfect!)
  • print the new fields - unit, value (to make the alpha-sort work properly it's zero-padded, fixed-length) and original line
  • index the last character of the size field
  • pull out the numeric portion of the size
  • sort the results, discard the extra columns

Try it without the cut command to see what it's doing.

Here's a version which does the sorting within the AWK script and doesn't need cut:

du -h |
   awk '{idx = sprintf("%s %08.2f %s", 
         index("KMG", substr($1, length($1))),
         substr($1, 0, length($1)-1), $0);
         lines[idx] = $0}
    END {c = asorti(lines, sorted);
         for (i = c; i >= 1; i--)
           print lines[sorted[i]]}'
2
  • thank you! this is the first example that works for me in OS X 10.6 not counting the perl/phython-scripts. and thanks again for the good explanation. always nice to learn something new. awk sure is a powerful tool.
    – Wolf
    May 4, 2011 at 12:09
  • Great thanks for that. I changed the du to du -sh * to show just the immediate files and directories without recursive descent.
    – HankCa
    Dec 9, 2016 at 22:01
20

Here's an example that shows the directories in a more compact summarized form. It handles spaces in directory/filenames.

% du -s * | sort -rn | cut -f2- | xargs -d "\n" du -sh

53G  projects
21G  Desktop
7.2G VirtualBox VMs
3.7G db
3.3G SparkleShare
2.2G Dropbox
272M apps
47M  incoming
14M  bin
5.7M rpmbuild
68K  vimdir.tgz
1
  • 1
    macOS/OSX users be warned that the mac version of xargs doesn’t support the -d flag, and if you omit it, any directories that contain a space have each word parsed separately which of course fails.
    – jasonology
    Jul 19, 2017 at 4:40
16

sort files by size in MB

du --block-size=MiB --max-depth=1 path | sort -n
11

I've a simple but useful python wrapper for du called dutop. Note that we (the coreutils maintainers) are considering adding the functionality to sort to sort "human" output directly.

2
  • 1
    +1 for one of the rare, valid exceptions to "do one thing and do it right". Unless someone gets sort to understand the SI-prefix and/or the binary prefixes. Mar 18, 2009 at 22:20
  • And as ptman mentions below: ta da! (new sort flag)
    – Tobu
    Jul 1, 2010 at 16:13
11

Got another one:

$ du -B1 | sort -nr | perl -MNumber::Bytes::Human=format_bytes -F'\t' -lane 'print format_bytes($F[0])."\t".$F[1]'

I'm starting to like perl. You might have to do a

$ cpan Number::Bytes::Human

first. To all the perl hackers out there: Yes, I know that the sort part can also be done in perl. Probably the du part, too.

10

This snippet was shameless snagged from 'Jean-Pierre' from http://www.unix.com/shell-programming-scripting/32555-du-h-sort.html. Is there a way I can better credit him?

du -k | sort -nr | awk '
     BEGIN {
        split("KB,MB,GB,TB", Units, ",");
     }
     {
        u = 1;
        while ($1 >= 1024) {
           $1 = $1 / 1024;
           u += 1
        }
        $1 = sprintf("%.1f %s", $1, Units[u]);
        print $0;
     }
    '
1
  • i think if it is a very big number, then the unit is gone and the number displayed is small... try 23423423432423 Apr 22, 2015 at 9:18
9

Use the "-g" flag

 -g, --general-numeric-sort
              compare according to general numerical value

And on my /usr/local directory produces output like this:

$ du |sort -g

0   ./lib/site_ruby/1.8/rubygems/digest
20  ./lib/site_ruby/1.8/rubygems/ext
20  ./share/xml
24  ./lib/perl
24  ./share/sgml
44  ./lib/site_ruby/1.8/rubygems/package
44  ./share/mime
52  ./share/icons/hicolor
56  ./share/icons
112 ./share/perl/5.10.0/YAML
132 ./lib/site_ruby/1.8/rubygems/commands
132 ./share/man/man3
136 ./share/man
156 ./share/perl/5.10.0
160 ./share/perl
488 ./share
560 ./lib/site_ruby/1.8/rubygems
604 ./lib/site_ruby/1.8
608 ./lib/site_ruby
1
  • 5
    That doesn't give the human-readable output, though, which is what the OP was looking for.
    – Jenn D.
    Feb 25, 2009 at 17:24
7

Found this one on line... seems to work OK

du -sh * | tee /tmp/duout.txt | grep G | sort -rn ; cat /tmp/duout.txt | grep M | sort -rn ; cat /tmp/duout.txt | grep K | sort -rn ; rm /tmp/duout.txt
1
6

Another one:

du -h | perl -e'
@l{ K, M, G } = ( 1 .. 3 );
print sort {
    ($aa) = $a =~ /(\w)\s+/;
    ($bb) = $b =~ /(\w)\s+/;
    $l{$aa} <=> $l{$bb} || $a <=> $b
  } <>'
6

I learned awk from concocting this example yesterday. It took some time, but it was great fun, and I learned how to use awk.

It runs only du once, and it has a output much similar to du -h

du --max-depth=0 -k * | sort -nr | awk '{ if($1>=1024*1024) {size=$1/1024/1024; unit="G"} else if($1>=1024) {size=$1/1024; unit="M"} else {size=$1; unit="K"}; if(size<10) format="%.1f%s"; else format="%.0f%s"; res=sprintf(format,size,unit); printf "%-8s %s\n",res,$2 }'

It shows numbers below 10 with one decimal point.

6

Here is the simple method I use, very low resource usage and gets you what you need:

du --max-depth=1 | sort -n | awk 'BEGIN {OFMT = "%.0f"} {print $1/1024,"MB", $2}'

0 MB ./etc
1 MB ./mail
2 MB ./tmp
123 MB ./public_html
5

du -cka --max-depth=1 /var/log | sort -rn | head -10 | awk '{print ($1)/1024,"MB ", $2'}

4

If you need to handle spaces you can use the following

 du -d 1| sort -nr | cut -f2 | sed 's/ /\\ /g' | xargs du -sh

The additional sed statement will help alleviate issues with folders with names such as Application Support

1
  • Just tried this on macOS Sierra. Works as expected. Nice!
    – jasonology
    Jul 19, 2017 at 4:44
2

http://dev.yorhel.nl/ncdu

command: ncdu

Directory navigation, sorting (name and size), graphing, human readable, etc...

1
  • 1
    Great utility, but not installed by default on any OS I'm aware of. Not necessarily a problem, but one more program to have to look after...
    – voretaq7
    Nov 29, 2011 at 23:07
2

Another awk solution -

du -k ./* | sort -nr | 
awk '
{split("KB,MB,GB",size,",");}
{x = 1;while ($1 >= 1024) 
{$1 = $1 / 1024;x = x + 1} $1 = sprintf("%-4.2f%s", $1, size[x]); print $0;}'


[jaypal~/Desktop/Reference]$ du -k ./* | sort -nr | awk '{split("KB,MB,GB",size,",");}{x = 1;while ($1 >= 1024) {$1 = $1 / 1024;x = x + 1} $1 = sprintf("%-4.2f%s", $1, size[x]); print $0;}'
15.92MB ./Personal
13.82MB ./Personal/Docs
2.35MB ./Work Docs
1.59MB ./Work Docs/Work
1.46MB ./Personal/Raa
584.00KB ./scan 1.pdf
544.00KB ./Personal/Resume
44.00KB ./Membership.xlsx
16.00KB ./Membership Transmittal Template.xlsx
2

Here is an example

du -h /folder/subfolder --max-depth=1 | sort -hr

Returns:

233M    /folder/subfolder
190M    /folder/subfolder/myfolder1
15M     /folder/subfolder/myfolder4
6.4M    /folder/subfolder/myfolder5
4.2M    /folder/subfolder/myfolder3
3.8M    /folder/subfolder/myfolder2

You could also add | head -10 to find the top 10 or any number of sub-folders in the specified directory.

1

Voilà:

du -sk /var/log/* | sort -rn | awk '{print $2}' | xargs -ia du -hs "a"
1

I had been using the solution provided by @ptman, but a recent server change made it no longer viable. Instead, I'm using the following bash script:

#!/bin/bash
# File: duf.sh
# list contents of the current directory by increasing 
#+size in human readable format

# for some, "-d 1" will be "--maxdepth=1"
du -k -d 1 | sort -g | awk '
{
if($1<1024)
    printf("%.0f KB\t%s",$1,$2);
else if($1<1024*1024)
    printf("%.1f MB\t%s",$1/1024,$2);
else
    printf("%.1f GB\t%s",$1/1024/1024,$2);
}'
1
  • The BSD du -d 1 syntax has been supported by GNU du since coreutils 8.6 was released in 2010 (though its first Red Hat availability was RHEL 7 in 2014), so you no longer need --maxdepth=1. I only found out about this recently myself.
    – Adam Katz
    May 3, 2017 at 21:47
1

du -s * | sort -nr | cut -f2 | xargs du -sh

1
  • That's not a great solution, because it traverses the file system twice.
    – Paul Gear
    Sep 30, 2015 at 0:06
1

There are a lot of answers here, many of which are duplicates. I see three trends: piping through a second du call, using complicated shell/awk code, and using other languages.

Here is a POSIX-compliant solution using du and awk that should work on every system.

I've taken a slightly different approach, adding -x to ensure we stay on the same filesystem (I only ever need this operation when I'm short on disk space, so why weed out stuff I've mounted within this FS tree or moved and symlinked back?) and displaying constant units to make for easier visual parsing. In this case, I typically choose not to sort so I can better see the hierarchical structure.

sudo du -x | awk '
  $1 > 2^20 { s=$1; $1=""; printf "%7sG%s\n", sprintf("%.2f",s/2^21), $0 }'

(Since this is in consistent units, you can then append | sort -n if you really want sorted results.)

This filters out any directory whose (cumulative) content fails to exceed 512MB and then displays sizes in gigabytes. By default, du uses a 512-byte block size (so awk's condition of 220 blocks is 512MB and its 221 divisor converts the units to GB — we could use du -kx with $1 > 512*1024 and s/1024^2 to be more human-readable). Inside the awk condition, we set s to the size so we can remove it from the line ($0). This retains the delimiter (which is collapsed to a single space), so the final %s represents a space and then the aggregated directory's name. %7s aligns the rounded %.2f GB size (increase to %8s if you have >10TB).

Unlike most of the solutions here, this properly supports directories with spaces in their names (though every solution, including this one, will mishandle directory names containing line breaks).

0

At least with the usual tools, this will be hard because of the format the human-readable numbers are in (note that sort does a "good job" here as it sorts the numbers - 508, 64, 61, 2, 2 - it just can't sort floating point numbers with an additional multiplier).

I'd try it the other way round - use the output from "du | sort -n -r" and afterwards convert the numbers to human-readable format with some script or program.

0

What you can try is:

for i in `du -s * | sort -n | cut -f2`
do
  du -h $i;
done

Hope that helps.

3
  • that's what xargs does ;-)
    – cadrian
    Feb 25, 2009 at 14:05
  • hehe, I always forget about xargs. ;) At the end of the day, whatever gets the job done imo.
    – Christian Witts
    Feb 25, 2009 at 15:05
  • MacOSX by default (i.e. outside of home-brew) doesn't support a proper xargs so this form was necessary. However for files with spaces in them you need to set IFS: IFS=$'\n'
    – HankCa
    Jan 30, 2016 at 11:16

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .