Regular Expression Capabilities

Individual Characters   Meta-characters Meta-character classes

Groupings Alternations Backreferences

Character classes Quantifiers

Individual characters

Meta-characters

Meta-characters which must be escaped
\	|	*	?
+	.	(	)
[	]	{	}

An example regular expression is 1+1. This does not mean 1+1 as it looks, because the + is a quantifier (see the section Quantifiers below). To make 1+1, you must escape the + in the regular expression, making the proper regular expression 1\+1.

Meta-character classes

Meta-character class	Characters included	Description
.	Any character in the current character universe (including \n)
\d	0123456789	Digits
\D	Any character in the current character universe, excluding the members of \d
\s	[SPACE][HT][VT][NL][FF]	Whitespaces
\S	Any character in the current character universe, excluding the members of \s
\w	ABCDEFGHIJKLMNOPQRSTUVWXYZ abcdefghijklmnopqrstuvwxyz0123456789_	Alphanumerics and _
\W	Any character in the current character universe, excluding the members of \w
\u{1}	ABCDEFGHIJKLMNOPQRSTUVWXYZ	Uppercase letters
\u{2}	abcdefghijklmnopqrstuvwxyz	Lowercase letters
\u{4}	0123456789	Digits
\u{8}	!@#$%^&*	Shift-with-numbers
\u{16}	;`:'[SPACE],".?_	Punctuation
\u{32}	(){}[]	Closures
\u{64}	~\/|	Others
\u{128}	+-=<>	Math
\u{NNN}	NNN is a number in decimal between 0 and 255, representing the sum of the pre-defined universe character class numbers. The resulting character class will be the union of all the included pre-defined character universes. Example: \u{233} 233 = 128 + 64 + 32 + 8 + 1 So, \u{233} will be the union of \u{1}, \u{8}, \u{32}, \u{64}, and \u{128}
\U{NNN}	NNN is a number in decimal between 0 and 255, representing the sum of the pre-defined universe character class numbers. The resulting character class will be any character in the current universe, excluding the members of the union of all the included pre-defined universe character classes. Example: \U{189} 189 = 128 + 32 + 16 + 8 + 4 + 1 So, \U{189} will be any character in the current character universe, excluding the members of the union of \u{1}, \u{4}, \u{8}, \u{32}, and \u{128}

Groupings

		> regldg -m 35 "(firstpart)anotherpart(second(third)part)" firstpartanotherpartsecondthirdpart

Alternations

		> regldg "ab|cd" ab cd

Alternations are often used with groupings when there are things in the regular expression which are not to be involved in the "this" or "that" game.

		> regldg "fla(t|pper)" flat flapper

regldg can also use multiple alternations to use "this" or "that" or "that" or "that" or "that".

		> regldg "(spl|th|fl|r)at" splat that flat rat

Backreferences

		> regldg -us 19 -m 46 "(Pat|Grandma) went to school today in \1's car\." Pat went to school today in Pat's car. Grandma went to school today in Grandma's car.

		> regldg -m 9 -us 19 "(a(b)c) \1 \2" abc abc b

regldg includes an alternative method to use backreferences. Instead of \1 to mean a backreference to grouping 1, you can use \!{1}. This will completely avoid the ambiguity of whether it is a backreference or an octally-specified character. (This is, of course, as long as you know this syntax. Otherwise, you might be completely confused as to what it is!) In action:

		> regldg -m 9 -us 19 "(a(b)c) \\!{1} \\!{2}" abc abc b

Note the double-\s... these were required for me to enter this regex in a tcsh. To avoid this problem, you could use the command line option --file=- and enter the regex directly into the program instead.

Character classes

		> regldg "[ab][cd]" ac bc ad bd

Some meta-characters don't need to be escaped while in character classes. These are (, *, +, ?, {, [, |, ), and }. The \ and . characters definitely need to be escaped in a character class. The range character - and the end-character-class character ] must be escaped unless they are the only character in the character class.

		> regldg -uc 0 "[(*+?{[|)}\\\-\]\.]" ( * + ? { [ | ) } \ - ] .

regldg is also capable of negated character classes, that is, character classes starting with the ^ character. A negated character class represents all characters in the current character universe, execpt those explicitly written in the negated character class.

		> regldg -us 2 "[^abcde]" f g h i j k l m n o p q ...

[-] and []] are both handled correctly: [-] is a character class containing only a - character, and []] is a character class containing only a ] character. Both are actually silly, because a one-element character class could instead be just that character. In any other character class, the - and ] characters are meta-characters, and need to be escaped.

Quantifiers

Quantifier	Meaning
*	The previous character, character class, meta-character class or group occurs between 0 to unlimited times (inclusive). (Unlimited is controlled by the maximum word length of the program.)
+	The previous character, character class, meta-character class or group occurs between 1 to unlimited times (inclusive). (Unlimited is controlled by the maximum word length of the program.)
?	The previous character, character class, meta-character class or group occurs between 0 to 1 time (inclusive).
{2}	The previous character, character class, meta-character class or group occurs 2 times.
{1,3}	The previous character, character class, meta-character class or group occurs between 1 to 3 times (inclusive).
{4,}	The previous character, character class, meta-character class or group occurs between 4 and unlimited times (inclusive). (Unlimitied is controlled by the maximum word length of the program.)

It is assumed that you have a good understanding of the usage of these items. A very important example, however, is when a quantifier acts on groups containing alternations. In a single word of output, any number of sides of the alternation could be used. The regular expression [a|b]{2} will produce aa, bb, AND ab and ba. Shown in an explicit example:

		> regldg "(ab|cd){2}" abab cdab abcd cdcd