dust Visual disk usage

Version 1.2.6 · source

Start here

dust
Biggest things under here
dust <dir>
Analyse a directory
dust -n 30
Show 30 entries
dust -d 2
Limit depth to 2
dust -r
Reverse the order
dust -X <name>
Ignore a name
dust -t
Group by file type
dust -j
JSON output

Concepts

  1. Walk dust <dir>
  2. Filter depth, size, regex
  3. Display tree with bars
  4. Pipe JSON to jq

Scope

dust <dir> <dir>
Analyse several directories
dust -d <n>
Show only n levels deep
dust -n <n>
Show n entries
dust -D
Directories only
dust -F
Files only: largest files
dust -x
Stay on one filesystem
dust -L
Follow symlinks
dust -i
Hide hidden files
dust -f
Count files instead of size

Filter

dust -X <name>
Ignore files or dirs by name
dust -e <regex>
Only entries matching regex
dust -v <regex>
Exclude entries matching regex
dust -z <size>
Hide entries below a size
Takes forms like 10M or 500KiB.
dust -t
Group by file type
dust --collapse <name>
Keep a folder collapsed

Display

dust -r
Reverse order: biggest at top
dust -p
Show full paths
dust -s
Apparent size, not disk use
dust -o <fmt>
Size unit: si, b, kb, kib, mb, gb, etc
dust -b
Hide bars and percentages
dust -B
Bars on the right
dust -c
No colours
dust -C
Force colours
dust --dim
Dim the bars
dust --skip-total
Omit the total row
dust -P
No progress indicator
dust -R
Screen-reader friendly output

Input and output

dust -j
Print JSON
dust --files-from <file>
Read paths, newline separated
dust --files0-from <file>
Read paths, NUL separated
dust -T <n>
Walker threads, useful on NFS

Recipes

What fills this disk
dust -x -d 2 /
Biggest files in a project
dust -F -n 20 <dir>
Skip noise, find big images
dust -X node_modules -X .git -e '\.(png|jpg)$' <dir>
Feed paths from find
find <dir> -name '*.log' -print0 | dust --files0-from -