awk¶

pattern-directed field processing (a small subset)
NAME
    awk - pattern-directed field processing (a small subset)

SYNOPSIS
    awk [-F c] 'program' [file]

DESCRIPTION
    Reads input a line (record) at a time, splits each into fields, and runs a
    program against it. Input is `file` if given, else standard input -- so it
    works in a pipe: `cat data | awk '{print $2}'`.

    The program is a single quoted argument (single or double quotes; awk strips
    them itself). It has the form

        [pattern] { action }

    pattern    /regex/ selects lines matching a basic regular expression
               (. * + ? ^ $, via the shared regex library); an empty pattern
               selects every line. A pattern with no action prints the line.
    action     { print items } prints a line; an empty action prints the whole
               record ($0).

    print items are comma-separated; a comma prints a space between them:
        $0             the whole record
        $1 .. $9       a field (by number)
        $NF            the last field
        NF             the number of fields on the record
        NR             the current record number (1-based)
        "text"         a string literal

    Fields split on runs of whitespace by default. -F c splits on the single
    character c instead (empty fields are kept).

OPTIONS
    -F c    use c as the field separator (default: whitespace)

EXAMPLES
    awk '{print $2}' data          print the 2nd field of each line
    awk '/error/{print}' log       print lines matching /error/
    awk '/error/' log              same (a bare pattern prints the line)
    awk '{print $1, $NF}' data     first and last field, space-separated
    awk -F: '{print $1}' pw        first colon-separated field
    awk '{print NR, $0}' data      number each line
    cat data | awk '{print $3}'    read from a pipe

LIMITATIONS
    A deliberately small subset: one rule per program, `print` only (no printf),
    no BEGIN/END, no variables or expressions, and no multi-file arguments.

SEE ALSO
    grep, sed, cut, cat