Bash Cut: How to Pull Out Portions of Text from Lines

The cut command reads lines from a file or standard input and prints specific portions based on fields, characters, or bytes. It works on one line at a time — no regex, no scripting — just fast column extraction at the terminal. If you’re new to working with text in the terminal, it’s one of the first tools worth learning.

Bash cut Command Syntax

The basic syntax takes an option and a file:

cut [OPTION] [FILE]

Without a file, cut reads from standard input, which makes it useful inside pipelines. Every invocation requires at least one of -f, -c, or -b. Calling cut without any option returns an error.

bash cut Command Flags and Options

Flag	Long Form	Purpose
-f	--fields	Select by field number (most common)
-d	--delimiter	Set a custom input delimiter (default: tab)
-c	--characters	Select by character position
-b	--bytes	Select by byte position
--complement	—	Print everything except the selected fields
-s	--only-delimited	Skip lines with no delimiter
--output-delimiter	—	Use a different character between output fields

The cut man page lists the full specification for each flag, including edge cases for byte ranges in multibyte locales.

How to Extract a Single Column with bash cut

Using the sample file example_data.txt:

Kai     Refsnes  30,Norway
Robin   Smith    25,Denmark
Sienna  Davis    40,Germany

Tab is the default delimiter, so pulling the first field needs nothing beyond -f1:

cut -f1 example_data.txt

Kai
Robin
Sienna

Setting a Custom Delimiter with -d

Tab separators don’t cover every case. CSV files, /etc/passwd, and many log formats use colons or commas. Pass -d with a single character to change the delimiter:

cut -d',' -f1 example_data.txt

Kai     Refsnes  30
Robin   Smith    25
Sienna  Davis    40

With a comma as the boundary, everything before the first comma counts as column one. The -d flag accepts exactly one character — passing two or more triggers the error cut: delimiter must be a single character.

Selecting Multiple or Range Columns

The -f flag accepts ranges and comma-separated lists. Use 1-2 for a range and 1,3 for non-consecutive fields:

cut -f1-2 example_data.txt

Kai     Refsnes
Robin   Smith
Sienna  Davis

To grab columns two through three:

cut -f2-3 example_data.txt

Refsnes  30,Norway
Smith    25,Denmark
Davis    40,Germany

Range notation follows a simple pattern: N for a single field, N-M for a span, N- for everything from N to the end of the line.

Using –complement to Exclude Fields

Instead of listing what you want, --complement lets you list what to drop. It prints every field except the ones specified:

cut --complement -f1 example_data.txt

Refsnes  30,Norway
Smith    25,Denmark
Davis    40,Germany

Column one is gone; everything else prints normally. The --complement flag is GNU-only and not available on macOS’s built-in BSD version of cut.

Extracting by Character Position

The -c flag works on character index rather than delimiters. It’s useful for fixed-width logs or data where columns aren’t separated by a character:

cut -c1-3 example_data.txt

Kai
Rob
Sie

For multibyte characters (UTF-8 text), use -b instead to target exact byte positions rather than character counts.

Using bash cut in Pipelines

Because cut reads from standard input, it connects naturally with other commands via pipes. Pulling just the usernames from /etc/passwd takes one line:

cut -d':' -f1 /etc/passwd

Chaining with sort and uniq removes duplicates and alphabetizes the output in a single pass. The bash shell usage statistics for 2026 confirm that bash remains the dominant shell for scripting, making fluency with tools like cut practically universal in shell environments.

cut -d':' -f1 /etc/passwd | sort | uniq

When the output delimiter should differ from the input one, --output-delimiter handles the conversion without needing sed:

cut -d',' -f1,3 --output-delimiter=':' example_data.txt

Common Errors and How to Fix Them

Error	Cause	Fix
cut: delimiter must be a single character	More than one character passed to -d	Use exactly one character, e.g. -d','
cut: fields and positions are numbered from 1	Zero used as a field number	Start field numbering at 1, not 0
Entire line printed instead of a field	Delimiter not present in that line	Add -s to skip lines without the delimiter
Unexpected output on macOS	--complement not supported in BSD cut	Use awk or install GNU coreutils via Homebrew

bash cut Quick Reference

Task	Command
First column (tab-separated)	cut -f1 file.txt
Comma-separated, column one	cut -d',' -f1 file.txt
Columns two through three	cut -f2-3 file.txt
All columns except the first	cut --complement -f1 file.txt
Characters 1 to 5 of each line	cut -c1-5 file.txt
Skip lines without the delimiter	cut -d':' -f1 -s file.txt
Change output delimiter	cut -d',' -f1,2 --output-delimiter=':' file.txt
Field from piped command	echo "a:b:c" | cut -d':' -f2

FAQs

What is the bash cut command used for?

The cut command extracts specific fields, characters, or bytes from each line of a file or input stream. It’s commonly used to parse CSV files, log files, and command output in shell scripts.

How do I use a custom delimiter with bash cut?

Use the -d flag followed by a single character, such as cut -d',' -f1 file.txt for comma-separated data. The default delimiter is a tab character.

What is the difference between -f, -c, and -b in cut?

-f selects by field using a delimiter, -c selects by character position, and -b selects by byte position. For ASCII files all three behave similarly, but -b and -c differ for multibyte characters.

Can I select non-consecutive fields with bash cut?

Yes. Use a comma-separated list with -f, such as cut -f1,3 file.txt to print the first and third fields. For a continuous range, use -f1-3.

How do I use bash cut with pipes?

Pipe the output of another command directly into cut, for example: cat /etc/passwd | cut -d':' -f1. No file argument is needed when reading from standard input.