Regular expression

A simple regular expression uses no special characters for defining the string to be used in a search it contains only a string you want you want to find in you test string. If we need to assign a pattern (rules) the test line should match we have to use special characters. For example, if we want to find out if my test string contains a phone number in the form ***-**** we need to use special character \d, which indicates that any digit matches this character but nothing but a digit, then my regular expression would be:

/\d\d\d-\d\d\d\d/;

matching metacharacters
Character	Matches	Example
\b	Word boundary	/\bor/ matches "origami" and "or" but not "normal" /or\b/ matches "traitor" and "or" but not "perform" /\bor\b/ matches "or" and nothing else
\B	Word nonboundary	/\Bor/ matches "normal" but not "origami" /or\B/ matches "normal" and "origami" but not "traitor" /\Bor|B/ matches "normal" but not "origami" or "traitor"
\d	Numeral 0 through 9	/\d\d\d/ matches "212" and "415" but not "B17" or "ABC"
\D	Nonnumeral	/\D\D\D/ matches "ABC" and "GEF" but not "B17" or "123"
\s	Single white space	/over\sbite/ matches "over bite" but not "overbite" or "over  bite"
\S	Single nonwhite space	/over\Sbite/ matches "over-bite" but not "overbite" or "over bite"
\w	Letter, numeral, or underscore	/A\w/ matches "A1" and "AC" but not "A+"
\W	Non letter, numeral, or underscore	/A\W/ matches "A+" but not "AC" and "A2"
\A	At the beginning of the string	/\AFread/ matches "Fred is OK" but not "I'm with Fred" or "Is Fred here?"
\Z	At the end of the string or before newline at the end	/Fread\Z/ matches "I'm with Fred\n and Bob" but not "Fred is OK" or "Is Fred here?"
\z	At the end of the string	/Fread\z/ matches "I'm with Fred" but not "Fred is OK" or "Is Fred here?"
.	Any character except new line	/.../ matches "abC", "12f", "1+ ", or ant three characters
[...]	Character set	/[AN]BC/ matches "ABC" and "NBC" but not "BBC"
[^...]	Negated character set	/[^AN]BC/ matches "BBC" and "CBC" but not "ABC" or "NBC"
Counting metacharacters
Character	Matches last character	Example
*	Zero or more times	/Ja*vaScript/ matches "JvaScript", "JavaScript", and "JaaaavaScript" but not "JuvaScript"
?	Zero or one time	/Ja?vaScript/ matches "JvaScript" or "JavaScript" but not "JaavaScript"
+	One or more times	/Ja+vaScript/ matches "JavaScript" or "JaaaavaScript" but not "JvaScript"
{n}	Exactly n times	/Ja{2}vaScript/ matches "JaavaScript" but not "JvaScript" or "JaaaavaScript"
{n,}	n or more times	/Ja{2,}vaScript/ matches "JaavaScript" or "JaaaavaScript" but not "JvaScript"
{n, m}	At least n at most m times	/Ja{2,3}vaScript/ matches "JaavaScript" or "JaaavaScript" but not "JvaScript" or "JaaaaavaScript"
positional metacharacters
Character	Matches located	Example
^	At the beginning of the string	/^Fread/ matches "Fred is OK" but not "I'm with Fred" or "Is Fred here?"
$	At the end of the string or before newline at the end	/Fread$/ matches "I'm with Fred\n and Bob" but not "Fred is OK" or "Is Fred here?"

For example if you want to make sure that a match for a Roman numeral is found only when it is at the start of a line and has a dot after it you check for the match

   /^[IVXMDC]+\./

Not to be confused with the metacharacters listed in the table above are escaped string characters for

Symbol	Escape symbol	Description
tab	\t	Tabulation symbol
newline	\n	New line symbol
return	\r	carriage return
formfeed	\f	Formfeed symbol (printer command)
vtab	\v	Vertical tabulation symbol
.	\.	Dot
^	\^	Caret symbol
$	\$	Dollar sign
\	\\	Backslash
/	\/	Slash
-	\-	Dash
(	\(	Open parenthesis
)	\)	Close parenthesis

Modifiers

• i - Do case-insensitive pattern matching.
• m - Treat string as multiple lines. That is, change "^" and "$" from matching the start or end of the string to matching the start or end of any line anywhere within the string.
• s - Treat string as single line. That is, change "." to match any character whatsoever, even a newline, which normally it would not match.
• x - Extend your pattern's legibility by permitting whitespace and comments.
• g - Perform a global search/substitute.

Using Regular Expressions

$str = "My phone number is 123-3445. This is my home phone.";
if( $str =~ /\d{3}-\d{4}/ ){
   print "There is a phone number in the string '$str'\n";
}
else{
   print "There is not a phone number in the string '$str'\n";
}

We also often need to know not only if there is match or not, but also what is the substring that matches the pattern. Perl provides several special variables for that purpose:

• $& - contains the matching substring
• $` - contains the substring on the left of the match
• $' - contains the substring on the right of the match

$str = "My phone number is 123-3445. This is my home phone.";
if( $str =~ /\d{3}-\d{4}/ ){
   print "There is a phone number in the string '$str'\n";
   print "The phone is: $&\n";
   print "      Before: '$`'\n";
   print "       After: '$''\n";
}
else{
   print "There is not a phone number in the string '$str'\n";
}

If we want to find all matches in a string we can use operator =~ in a loop and setting the value of the variable $str to the substring on the right of the match:

$str = "My phones: 123-3456 (home), 234-4557 (office), 456-4564 (cell).";
while( $str =~ /\d{3}-\d{4}/ ){
   print "The phone is: $&\n";
   $str = $';
}

$str = "My phones: 123-3456 (home), 234-4557 (office), 456-4564 (cell).";
@phones = $str =~ /\d{3}-\d{4}/g;
foreach $phone (@phones){
   print "$phone\n";
}

In addition to operator =~ Perl has operator !~ that returns true if there is no match and false otherwise.

Getting information about a match

For example, if we are checking that a date was entered in either in "mm/dd/yyyy" or "mm-dd-yyyy" format and also need to know the values of the month, day, and year we can use the following regular expression:

$today = "1/24/2003";
if( $today =~ /\b(1[0-2]|0?[1-9])[\-\/](0?[1-9]|[12][0-9]|3[01])[\-\/]((19|20)\d{2})/ ){
   print "Date: $&\n";
   print "Month: $1\n";
   print "Day: $2\n";
   print "Year: $3\n";
   print "Century: $4\n";
}

• 1[0-2]|0?[1-9] for the month (can start with 1 and then have only 0, 1, or 2 as the second digit, or can start with 0 and have anything but zero as the second digit). We also put parenthesis around this part of the regular expression to extract month from the matching string.
• 0?[1-9]|[12][0-9]|3[01] for the day (can start with 0 and have at least 1 after that, can start with 1 or two and have any digit after that, or can start with 3 and have only 0 or 1 after). This part is also inside parenthesis because we will need to know the day of the month as well.
• (19|20)\d{2} for the year (can start with 19 or 20 and have exactly two digit after that). This part of the pattern is inside the parenthesis (that'll give us the year) and also has it's own part parenthesised. This subpart will give us the century.

String Replacement

$card = "6432 23452342-2342";
if( $card =~ /(\d\d\d\d)[\-\s]?(\d\d\d\d)[\-\s]?(\d\d\d\d)[\-\s]?(\d\d\d\d)/ ){
   	$card = "$1-$2-$3-$4";
}
else{
  	$card = "Invalid credit card number!";
}

To replace a part of a string that matches a regular expression we can use s/// regular expression. This expression includes a pattern (goes between the first and second slash) and a string to substitute with (goes between the second and the third slash). Operator =~ performs the substitution if used with s/// regular expression. Thus, the following example substitutes the first 4 digits in a credit card number with stars:

$card = "6432 23452342-2342";
if( $card =~ s/\d{4}/****/ ){
   print "$card\n";
}
else{
   print "Sorry, there is no match\n";
}

$card = "6432 23452342-2342";
$card =~ s/(\d\d\d\d)[\-\s]?/$1-/g; # separate 4-digit groups with dashes
$card =~ s/-$//;                    # remove the trailing dash
$card =~ s/\d{4}-/****-/g;          # substitute 4-digit groups with 4 stars
print "$card\n";

Please consult Perl regular expression documentations for more details.