Skip to main content

Disk usage summary, showing large dirs/files

Project description

dutree shows a summary of the directories/files which take up the most space.

Instead of showing only the root of the files with sizes, or the details of every file, it shows only the paths taking up the most space.

Example usage:

$ dutree /srv --apparent-size

Annotated output, where only paths of >5% of the total size are shown (which is about 4GB for this dataset):

12.1 G  /srv/data/audiofiles/
         ^-- audiofiles contains files/dirs with a total of 12.1G
             but it does NOT contain a single dir or file larger
             than 4G.
 4.3 G  /srv/data/callrecordings/unmatched/
 4.5 G  /srv/data/fax/
17.5 G  /srv/data/playlists/
34.4 G  /srv/data/twinfield_invoices/1/
 7.8 G  /srv/data/*
          ^-- data contains more files/directories than shown above
              but those that don't total above 4G are merged into
              this "leftover" node. that includes everything in
              /twinfield_invoices/ except for the /1/ subdirectory
              which has its individual listing above.
32   B  /srv/*
          ^-- only /data/ is in /srv/, but the directory itself also
              takes up a tiny bit of space
80.6 G  TOTAL (86558511658)

NOTE: The directories do not count the size of themselves, only of their contents. This explains any discrepancies with du -sb output.

Library usage:

>>> from dutree import Scanner
>>> scanner = Scanner('/srv')
>>> tree = scanner.scan(use_apparent_size=True)
>>> tree.app_size()

>>> len(tree.get_leaves())

>>> leaf0 = tree.get_leaves()[0]

>>> leaf0.app_size() / (1024.0 * 1024 * 1024)


  • Please run the python2 and python3 tests after developing:

    python2 dutree/
    python3 dutree/


  • v1.8

    • Add experimental feature to do scanning from a file instead of the filesystem. See commit bbebcf98 for details. Original code by Seppe Lucas.

  • v1.7

    • Skip /proc and /sys filesystems by default.

    • Use –xdev to stay on the same filesystem.

    • Uses –count-blocks by default; use –apparent-size to negate.

  • v1.6

    • Fix so the tests work with Python 3 as well.

    • Fix grave bugs with real size. The real size calculation was wrong sometimes, it could raise assertions or the leaf count could be off.

  • v1.5

    • Add apparent vs. real size. Deprecates node.size(). Use node.app_size() instead. Get the real used size using node.use_size(). Group by real used size by passing use_apparent_size=False to scan().

    • Don’t die if we cannot enter directories.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

dutree-1.8.tar.gz (15.3 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page