Home / ubuntu / Questions / 80244
Accepted
UAdapter
Asked: 2011-11-18 09:02:59 +0800 CST

How do I sort by human readable sizes numerically?

  • 772

for example I have command that shows how much space folder takes

du folder | sort -n

it works great, however I would like to have human readable form

du -h folder

however if I do that than I cannot sort it as numeric.

How to join du folder and du -h folder to see output sorted as du folder, but with first column from du -h folder

P.S. this is just an example. this technique might be very useful for me (if its possible)

bash

8 Answers

  1. Try something like:

    du -h folder | sort -h

    Alternatives: -n for Numerical sorting

    Note: the -h option of sort only exists in newer versions of Ubuntu.

    • 42
  2. Best Answer

    Here is a more general approach. Get the output of du folder and du -h folder in two different files.

    du folder > file1
du -h folder > file2

    The key part is this: concatenate file1 and file2 line by line, with a suitable delimiter.

    paste -d '#' file1 file2 > file3

    (assuming # does not appear in file1 and file2)

    Now sort file3. Note that this will sort based on file1 contents and break ties by file2 contents. Extract the relevant result using cut:

    sort -n -k1,7 file3 | cut -d '#' -f 2

    Also take a look at man sort for other options.

    You may also save this as an alias, for later re-use. To do so, add the following to the bottom of ~/.bashrc: 

    sorted-du () {
    paste -d '#' <( du "$1" ) <( du -h "$1" ) | sort -n -k1,7 | cut -d '#' -f 2
}

    Then, open a new terminal session and execute your new alias:

    sorted-du /home
    • 13

  3. This answer is valid for 10.04.4LTS and lower versions of Ubuntu.

    Unfortunatly the accurate answer which sorts K M G is difficult and complex:

    You can alias the entire du command with one that sorts human readable using this

    alias duf='du -sk * | sort -n | perl -ne '\''($s,$f)=split(m{\t});for (qw(K M G)) {if($s<1024) {printf("%.1f",$s);print "$_\t$f"; last};$s=$s/1024}'\'

    which I found here

    http://www.earthinfo.org/linux-disk-usage-sorted-by-size-and-human-readable/

    just cd into the folder you would like to know then duf

    you could add this duf alias to the end of your /home/user/.profile to make the duf command semi-permenant

    results:

    user@hostname:~$ duf
0.0K  Documenten
0.0K  Muziek
0.0K  Openbaar
0.0K  Sjablonen
0.0K  Video's
4.0K  backup_db.sql.g
4.0K  examples.desktop
12.0K xml printer ticket
52.0K hardinfo_report.html
152.0K    librxtxSerial.so
2.7M  jpos
4.4M  nxclient_3.5.0-7_amd64.deb
6.4M  nxnode_3.5.0-4_amd64.deb
6.8M  Downloads
7.4M  nxserver_3.5.0-5_amd64.deb
12.4M NetBeansProjects
18.1M mysqlworkbench.deb
28.3M Afbeeldingen
45.8M ergens-20110928-18.sql.gz
60.5M 2012-06-02ergens_archive.tar.gz
65.5M 2012-08-26ergens_archive.tar.gz
65.6M 2012-08-28ergens_archive.tar.gz
65.6M 2012-08-29ergens_archive.tar.gz
65.7M 2012-08-30ergens_archive.tar.gz
113.0M    Bureaublad
306.2M    ergens-20110928-18.sql

    Here is why du -sch /var/* | sort -n does not work see the sorting of MKKMMKKMMK

    user@hostname:~$ du -sch /var/* |sort -n

0 /var/crash
0 /var/local
0 /var/lock
0 /var/opt
8,0M  /var/backups
12K   /var/games
16K   /var/tmp
17M   /var/log
68M   /var/cache
104K  /var/spool
144K  /var/run
351M  /var/lib
443M  totaal
704K  /var/mail
    • 4

  4. Command GNU sort has the following option:

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

    To have this option on BSD/OSX, you can install sort from coreutils (via brew) and add the bin folder to your PATH into your rc files.

    So the command would looks like:

    du -ah . | sort -rh | head -20
    • 3

  5. For recent versions of Ubuntu, use du -h directory | sort -h .

    I use a form of this all the time for finding out-of-control files.

    • 1

  6. Open your terminal with Ctrl+Alt+t and type:

      du -sk * | sort -rn
    • 0

  7. The one I found working on all systems regardless of the version (sort -h doesn't work for me) is 

    du -h $(du * | sort -n | awk '{print $2}')

    the awk '{print $2}' basically prints file names.

    • 0

  8. This one handles filenames with whitespace or apostrophes, and works on systems which do not support xargs -d or sort -h:

    du -s * | sort -n | cut -f2 | tr '\n' '\0' | xargs -0 -I {} du -sh "{}"

    which results in:

    368K    diskmanagementd
392K    racoon
468K    coreaudiod
472K    securityd
660K    sshd
3.6M    php-fpm
    • 0