cut — Extract Fields, Characters and Bytes from Text Lines

cut is the lean tool for slicing exactly the column you need out of every line – whether that is a field in a CSV file, a character range in a log line, or a byte offset in fixed-width data. You tell it whether to cut by fields (-f), characters (-c) or bytes (-b), and cut returns the rest line by line. By default it splits fields on the tab character; with -d you set your own delimiter. In pipelines, cut is often the fastest answer when awk would be overkill.

Field Extraction

cut -f <n> <file> — Extract the nth field (tab-delimited by default).

cut -f 2 data.tsv

cut -f <n> -d '<delim>' <file> — Extract the nth field using a custom delimiter.

cut -f 1 -d ',' users.csv

cut -f <n>,<m> -d '<delim>' <file> — Extract multiple specific fields.

cut -f 1,3 -d ':' /etc/passwd

cut -f <n>-<m> -d '<delim>' <file> — Extract a range of fields.

cut -f 2-4 -d ',' data.csv

cut -f <n>- -d '<delim>' <file> — Extract from field n to the end.

cut -f 3- -d ':' /etc/passwd

cut -f -<n> -d '<delim>' <file> — Extract from the first field up to field n.

cut -f -3 -d ',' data.csv

cut -f <n> -d '<delim>' -s <file> — Suppress lines that do not contain the delimiter.

cut -f 2 -d ',' -s mixed.txt

cut -f <n> -d '<delim>' --output-delimiter='<sep>' — Use a different delimiter for the output.

cut -f 1,3 -d ':' --output-delimiter=' | ' /etc/passwd

Character Extraction

cut -c <n> <file> — Extract the nth character from each line.

cut -c 1 file.txt

cut -c <n>-<m> <file> — Extract a range of characters.

cut -c 1-10 file.txt

cut -c <n>,<m>,<o> <file> — Extract specific character positions.

cut -c 1,5,10 file.txt

cut -c <n>- <file> — Extract from character n to the end of each line.

cut -c 5- file.txt

cut -c -<n> <file> — Extract the first n characters of each line.

cut -c -20 file.txt

Byte Extraction

cut -b <n> <file> — Extract the nth byte from each line.

cut -b 1 file.txt

cut -b <n>-<m> <file> — Extract a range of bytes (useful for fixed-width binary data).

cut -b 1-16 binary.dat

cut -b <n>- <file> — Extract from byte n to the end.

cut -b 5- data.bin

Complement

cut --complement -f <n> -d '<delim>' <file> — Extract all fields EXCEPT the specified ones.

cut --complement -f 2 -d ',' data.csv

cut --complement -c <n>-<m> <file> — Remove a range of characters, keep everything else.

cut --complement -c 1-5 file.txt

Common Patterns

cut -f 1 -d ':' /etc/passwd — Extract all usernames from /etc/passwd.

cut -f 1 -d ':' /etc/passwd

echo $PATH | cut -f 1 -d ':' — Get the first directory in the PATH variable.

echo $PATH | cut -f 1 -d ':'

cut -f 1,3 -d ':' /etc/passwd | sort — Extract username and UID, then sort.

cut -f 1,3 -d ':' /etc/passwd | sort -t ':' -k 2 -n

ps aux | cut -c 1-80 — Truncate process listing to 80 characters wide.

ps aux | cut -c 1-80

<command> | cut -f <n> -d ' ' — Extract a column from space-delimited command output.

df -h | cut -c 1-20,45-52

rev <file> | cut -f 1 -d '.' | rev — Extract file extensions by reversing, cutting, and reversing again.

ls | rev | cut -f 1 -d '.' | rev

head -1 data.csv | cut -f 1- -d ',' --output-delimiter=$'\n' — List CSV column headers one per line.

head -1 data.csv | cut -f 1- -d ',' --output-delimiter=$'\n'

Conclusion

cut wins through simplicity: when your data has a clear, consistent delimiter, you extract the column you want faster and more readably than with any other tool. Mind its limits, though – cut knows only a single-character delimiter (it cannot treat repeated spaces as one, the way awk does) and cannot reorder fields. For multiple consecutive spaces, squeeze them first with tr -s ' '; as soon as logic, conditions or variable separators come into play, awk is the better choice. With multibyte (UTF-8) characters, -c and -b differ – cut by character, not by byte, in that case.