cut(1)

cut – cut out selected portions of each line of a file

cut is a general purpose text splitter and extractor. It operates in two modes. The -d sets the delimiter and -f specifies the “fields” to exact. The -c option specifies the character positions to take. cut is really useable in day to day shell programming because you can use to take semi structured data (such as a person’s name) and extract structured bits (such as the given or family name). cut has one confounding quirk: Indexes start at 1 unlike most things in programming. Here are some examples:

$ echo "Adam Hawkins" | cut -d ' ' -f 1
Adam
$ echo "Adam Hawkins" | cut -c 1-4
Adam
$ echo "Adam Andrew Hawkins" | cut -d ' ' -f 1,2
Adam Andrew

Use Case

cut works perfectly for simple text manipulation and extraction like pulling data from CSV files, TAB separated data, or semi structured text. It’s analogous to Ruby’s String#split and array accessors like foo[1..3]. It’s easier than awk (and more readable) for simple things. cut works with single character delimiters, so you cannot use a regex or other more complex splitting criteria. Consider extracting the path from a URI. Using cut here would be clunky because paths themselves may contain / (the delimiter character). sed or grep fit better in such cases because they support regular expressions.

Signature & Key Options

cut -b list [-n] [file ...]
cut -c list [file ...]
cut -f list [-d delim] [-s] [file ...]

This usage synopsis comes from the BSD version because it identifies the three different usage modes: bytes, characters, and delimiters. Remember that list indexes start at 1, not zero (Cut apologizes in advance for making you remember that).

• -b specifies byte list from file. Use this when you need to extract bytes from structured binary files (some TCP dump for instance). Combine with -n for multi-byte characters.
• -c select character list from file. Specify multiple ranges with ,. Examples: take 5 characters - -c 5; take the first 3 and 10-15; -c 1-4,10-15.
• -f selects list fields from file split by -d. -d defaults to \t. Specify multiple ranges with ,. Examples: second word from a sentence -echo "the quick brown fox" | cut -d ' ' -f ; take the first and third words - echo "the quick brown fox" | cut -d ' ' -f 1,3. -d only accept single character values. -s surpasses lines without delimiter. Use this when input files may contain junk data.

GNU vs BSD

The BSD and GNU versions are similar functionality similar.

• The GNU version supports a -z or --zero-terminated useful when used with find -print0 and xargs -0.
• The GNU version supports --complement to invert the match specified with -b, -c, and -f. This option is useful when you have many fields and want to print all but a few of them