10 cut Command Examples in Linux (Extract Text Like a Pro)

cut reads from a file or standard input and always writes to standard output — it never modifies the source file. To save extracted data, redirect the output: cut -d ',' -f2 data.csv > names.txt. For more on output redirection see the Linux I/O Redirection Cheat Sheet.

You can select non-contiguous byte positions using a comma-separated list. For example, cut -b 1,3,5 report.txt outputs only bytes 1, 3, and 5 of each line. You can also select from a position to the end of the line with an open-ended range: cut -b 6- report.txt outputs everything from byte 6 onwards.

Fields in cut are numbered from 1, not 0. So -f1 is the first column, -f2 the second, and so on. This is different from most programming languages that use zero-based indexing. If you need to add a header row to the output, pipe the result through standard Linux commands such as sed.

Open-ended ranges work too. -f3- selects field 3 through the last field on every line, even when the number of fields varies between records. This is particularly useful for log files where the final column is a free-text message of variable length — you can safely capture the entire tail of each line.

Pair fields 1 and 7 to see each username alongside its assigned shell: cut -d ':' -f1,7 /etc/passwd. The output is formatted as username:/bin/bash, giving you a clear and scannable picture of which accounts have interactive shell access on the server.

--complement is a GNU cut extension and is available on all major Linux distributions (RHEL, CentOS, Ubuntu, Debian, Arch). It is not supported on the BSD cut that ships with macOS. If you are on macOS and need to drop a column, use awk with a conditional skip: awk -F',' 'BEGIN{OFS=","}{$2=""; gsub(/,,/,","); print}' roster.csv.

The cut command in Linux is used to extract specific sections from each line of a file or from standard input piped from another command. You can extract data by raw byte position (-b), by character position (-c), or by field number when the input is structured with a consistent delimiter (-d combined with -f). Common real-world uses include pulling columns from CSV files, extracting usernames from /etc/passwd, isolating error codes from log entries, and preprocessing data for import into databases or spreadsheets.

On files that contain only standard ASCII characters — which covers the vast majority of log files, CSV files, and system configuration files — -b and -c behave identically because each character is exactly one byte. The difference surfaces with multi-byte UTF-8 characters such as accented letters (é, ñ), Chinese or Arabic characters, or emoji. The -b flag counts raw bytes, which can cut through the middle of a multi-byte codepoint and produce garbled output. The -c flag is locale-aware and counts whole characters, keeping every codepoint intact. As a rule, always use -c when there is any possibility the file contains non-ASCII text.

Pass a comma-separated list of field numbers to -f to select non-adjacent columns: cut -d ',' -f1,3,5 data.csv extracts fields 1, 3, and 5. For a consecutive range, use a hyphen: -f2-4 extracts fields 2, 3, and 4. You can also mix both notations in a single expression: -f1,3-5,7 selects field 1, fields 3 through 5, and field 7. Remember that fields are numbered from 1, not 0, and cut always outputs them in their original left-to-right order regardless of how you list them.

The most common causes are: (1) the actual delimiter in your file does not match what you passed to -d — run cat -A yourfile.txt to reveal invisible characters (a ^I indicates a tab, which is the default delimiter when -d is omitted); (2) your data uses a multi-character separator, which cut does not support — switch to awk -F in that case; (3) there are leading or trailing spaces around the delimiter that prevent the field from being recognised as a clean match.

No, never. cut is a read-only tool — it reads from its input and writes to standard output, leaving the source file completely untouched. To save the extracted data to a new file, use output redirection: cut -d ',' -f2 roster.csv > names.txt. To replace the original file you would need to write to a temporary file first and then rename it, or use a tool like sed -i — but that is rarely the right approach for straightforward column extraction.

Use awk instead of cut when: (1) your delimiter is more than one character, (2) fields are separated by variable amounts of whitespace, (3) you need to reorder or combine columns in the output, (4) you need conditional logic (print only rows where field 4 exceeds 80000), or (5) you need arithmetic on field values. For simple, single-pass column extraction from consistently delimited files, cut is faster to type, easier to read, and slightly more efficient. For complex transformations, awk is the better tool.