regex

Regular Expressions

Basic Expressions

The simplest regular expressions are alphanumeric strings, e.g., /pattern/, will match the substring "pattern".

Wild Cards and Quantifiers

The character . matches any character, so for example, the pattern /.ed/ will match all of "sed" and "bed" and "1ed".

Quantifiers match the preceding expression multiple times. There are several quantifier expressions:

• * matches the preceding expression 0 or more times
• ? matches the preceding expression 0 or 1 times
• \{n\} matches the preceding expression n times, where n is a number
• \{m,n\} matches the preceding expression m to n times, where n is a number.

Combine wild cards with quantifiers to match arbitrary strings, e.g., /.*ed/ matches all of "sed", "ed", and "foo bar baz ed".

Grouping

You can cause an regular expression to be treating as one thing for quantifiers by surrounding it with \(...\). For example /\(foo\)\{3\}/ matches 'foofoofoo'.

Character Sets

Sometimes . is too powerful a wildcard, and what you really want is to match some characters but not others. Character sets let you do that. A character set is composed of square brackets around the characters you want to match.

For example, [AEIOU] matches any uppercase vowel. You can also invert the character set with a ^ after the [, i.e., [^AEIOU] matches anything BUT an uppercase vowel. For convenience, you may also list character ranges, e.g., [A-Z] or [0-9].

Beginning and End of a String

You can match the start and end of a string with ^ and $, respectively. As we will see with sed, regular expressions are often evaluated line by line over a file, and in these instances ^ and $ refer to the start and end of the line.

Alternation

You can match one or more alternatives with \|. For example /foo|bar/ matches both foo and bar. If you only want part of the regular expression to have alternatives, you can use \(...\). For example /\(Li\|U\)nix/ matches both Linux and Unix.

Programs other than sed

The above guide describes POSIX Basic Regular Expressions which are used by sed. Many regular expression tools, such as egrep and Python's re module Links to an external site., use a syntax based on "extended" regular expressions. In this style of regular expression, there is no backslash before |, (, ), {, or }, so one must write /(Li|U)nix/ instead of /\(Li\|U\)nix/. Instead, putting a backslash before a nonalphanumeric character always matches that character only.

Although there are standards for regular expressions, in practice, there is a lot of minor variation in the features and syntaxes that regular expression libraries support, so you should try to consult documentation for the tool you are using.