Perl regex

Navbar

computing: programming: Perl: regex

Overview

This article explains regular expressions in terms understandable to mere mortals, and also how to use them in Perl.

Related Articles

• regex: manpage documentation

Details

Special characters in regex:

• . = any character except newline (\n) (if /s option is included, then \n is also matched)
• * = 0 or more of previous character
• ^ = following string begins the line (except [^...] means "not these characters")
• $ = preceding string ends the line
• [] = list of characters which can satisfy the match at this position
  • Note: cannot use special characters like "." as part of list
• {} = # of repetitions of previous character:
  • {x} -> exactly x repetitions
  • {x,y} -> minimum of x repetitions, maximum of y repetitions
• | = alternatives
• + = 1 or more of previous character
• ? after +, *, or {} indicates non-greedy behavior, i.e. match the fewest characters, not the most
  • Apparently it can also mean "either zero or one of the preceding group or character", e.g. "the?" matches both "th" and "the"
• a-b = range of characters from a to b, e.g. "t-w" means any of t,u,v,w in that position
• ?= = lookahead (need explanation of how this works) a(?=b) returns "a, but only if it's followed by a b"; the a becomes part of the matched sequence, but the b does not
• ?<= = reverse lookahead (need explanation of how this works)

Operators used to invoke regex:

• =~ returns TRUE if pattern matches
• !~ returns FALSE if pattern matches
• m/ searches a string for a pattern match; returns true/false scalar and an array of matches (if () are used)
  • c don't reset pos on failed matches when using /g
  • g (Global) all occurrences – repeat the pattern search until there are no more matches
  • i case-Insensitive
  • m Multiline mode - ^ and $ match internal lines
  • o compile pattern Once
  • s match as a Single line - . (dot) matches \n
  • x eXtended legibility - free whitespace and comments
• s/pattern/replacement/gi; replaces pattern with replacement
  • all m/ options are available
  • e Evaluate replacement as an expression. May be specified multiple times. 'replacement' is interpreted as a double quoted string unless a single-quote (') is the delimiter.
• y/searchlist/replacelist/d: replaces each character found in searchlist with the corresponding character in replacelist
  • d just deletes matching characters
• tr/ is the same as y/

Examples

These examples have been tested only briefly.

• Replace "thingy" with "stuffs" in $string:
  • $string =~ s/thingy/stuffs/;
• Keep only the part of $string before the final "/" (using "|" as the delimiter instead of "/"):
  • $string =~ s|(.*)/[^/]*|$1|;
• ...after the final "/":
  • $string =~ s| ^.*/([^/]*)$|$1|;
• ...before the final "-":
  • $string =~ s|(.*)-[^-]*|$1|;
• ...before the final ".":
  • $string =~ s|(.*)\.[^\.]*|$1|;
• ...after the final "." (both of these return the full string if no "." is found):
  • $string =~ s|^.+\.(.+$)|$1|;
  • $string =~ s|^.*\.([^\.]*)$|$1|;

Links

Reference

• Regular Expressions (Perl-Compatible) in PHP
  • PHP also supports POSIX regex via the ereg*() series of functions, but these are apparently deprecated; to convert code from ereg* to PCRE, just add slashes around the string: /regex expression/

Articles

• 2007-10-24 A graphical tracer for Perl 6 regexes based on PCR: code for a graphical regex tracer written in Perl 6, with links to some demo sites