The cut command in Linux
cut extracts sections from each line of input: fields split by a delimiter (-d -f) or fixed character positions (-c). For simple, consistently delimited data it is faster to type than awk; the moment separators become variable whitespace, switch to awk.
How cut works
cut applies one fixed slicing rule to every line — fields split on a single literal character, or absolute character positions — and outputs the selected pieces. No regex, no state, no surprises: a pure per-line transformation, which is why it is fast and why it composes so cleanly in pipelines.
Its limits are the flip side of that purity. One literal delimiter character means variable whitespace defeats it (that is awk territory). Selected fields always emerge in original order — -f3,1 will not reorder. And quoted CSV, where a comma may be data, is beyond any line-slicing tool: the escaping rules require a real parser. Knowing precisely where cut ends is what lets you reach for it with confidence inside its territory.
Syntax
cut -d DELIM -f FIELDS [FILE] | cut -c RANGES [FILE]Common options
| Option | What it does |
|---|---|
-d"," | Field delimiter (a single character). |
-f 1,3 | Which fields to keep (lists and ranges: 1,3 or 2-4 or 3-). |
-c 1-8 | Character positions instead of fields. |
--complement | Keep everything except the named fields. |
How to use cut: examples
$ cut -d',' -f1,3 users.csvColumns 1 and 3 of a comma-separated file.
$ cut -d':' -f1 /etc/passwdUser names from the system account file.
$ cut -c1-10 access.logThe first ten characters of each line — e.g. a fixed-width date.
$ echo 'a:b:c:d' | cut -d':' -f2-Field 2 to the end.
$ history | cut -c8-Strip the line numbers from your shell history output.
Real-world use cases for cut
Extracting fields from system files
All usernames: cut -d: -f1 /etc/passwd. All shells in use: cut -d: -f7 /etc/passwd | sort -u. Colon-separated system files are cut's home turf — single-character delimiter, fixed positions, zero surprises.
Trimming fixed-width logs
Log lines starting with a fixed 19-character timestamp: cut -c21- strips it for cleaner diffing, and cut -c1-10 keeps just the date for grouping with sort | uniq -c.
Pro tips and common mistakes
- Field ranges compose: -f1,3-5 takes field 1 plus 3 through 5; -f2- takes everything from field 2 on.
- cut cannot reorder: -f3,1 still outputs in file order (1 then 3). Reordering is awk territory: {print $3, $1}.
- Lines without the delimiter pass through untouched; add -s to drop them instead.
- Real-world CSVs with quoted commas defeat cut and awk alike — use a real parser (python csv, jq for JSON) when data has escaping.
Frequently asked questions about cut
Why does cut print the whole line unchanged?
The delimiter was not found on that line. cut passes such lines through untouched by default; add -s to suppress them.
Can cut use a multi-character delimiter?
No — -d takes exactly one character. For multi-character separators or variable whitespace, use awk -F or sed.
How do I cut on spaces when columns are aligned?
Aligned columns usually mean variable numbers of spaces, which cut can't handle. Use awk '{print $2}' — awk collapses whitespace automatically.
Related commands
Stream-edit text: substitute, delete lines, and edit files in place with -i.
awkPull columns out of any output, filter rows, and compute sums — the shell's spreadsheet.