Regular Expressions: Concepts

1
Regular Expressions Concepts

• CSC492 Topics in Perl
• Joe Reynoldson
• 3/14/05

2
Regular Expressions (regex) Basics

• The purpose of a regex to match a string to a
  pattern
• The syntax of regex is simple (although sometimes
  confusing), and it basically represents a whole
  new language
• Regex is not unique to Perl, but regex languages
  are unique
• Many (including Java) imitate perl regex
• Regexes and globs are 2 different things
• Globs are wildcard matches, often in UNIX or DOS
  shell
• Example .pl

3
Regex explained

• Regexes divide an infinite set of input strings
  into 2 groups
• Matched
• Didn't match
• There's no in between (fuzzy matching)
• regexes are often used as boolean statements in
  conditional expressions (such as in if statements
  or while loops)

4
Matching with the matching operator

• The matching operator ( m// ) matches _ against
  the pattern specified between the slashes
• Most Perl hackers leave off the leading 'm' and
  just use slashes

while(ltINgt) if( /Waldo/ ) print "I
found Waldo in _!\n"
5
Metacharacters Wildcard ( . )

• There are many characters which have special
  meaning in regexes
• The dot ( . ) is a wildcard which matches any
  character (except newline... we'll get to that)
• Matches if _ is I have a cat or I have a car
• Also matches can, cab, cam, and ca9

/I have a ca./
6
Metacharacters Escape ( \ )

• To match a period, you must escape it
• The escape character ( \ ) is our second
  metacharacter
• Use escape whack to match a whack/\\/
• Use whack-a-mole to relieve tension

/3\.14159/
7
Matching Operator Variation

• The slashes can be replaced by pretty much any
  character
• This is handy when matching against data with
  many slashes (such as the full path to a file)
• Perhaps m/home/j/jreynold would be more
  explicit for non-hackers
• Better than the alternative

if( /home/j/jreynold ) print "_ contains
my home directory\n"
/\/home\/j\/jreynold/
8
A Quick CGI Example

• Suppose you want to determine if a web page was
  requested by a browser in the usd.edu domain
• Of course this regex fails miserably, but how?
  (Hint There are an infinite number of input
  strings)

_ ENV'REMOTE_HOST' if(/\.usd\.edu/ )
print "You appear to be on a ltbgtU.lt/bgt
computer\n"
9
Metacharacters Simple Quantifiers

• Quantifiers allow a programmer to specify how
  many matches s?he wants (hee hee)
• Here they are in no particular order
• ? matches a pattern 0 or 1 time (/s?he/ would
  match she or he, and now youre in on my geeky
  joke! You too can be the life of the party!)
• matches the pattern 0 or more times (/she/
  matches he, she, sshe, ssshe, )
• matches the pattern 1 or more times (/she/
  matches she, sshe, ssshe, )

10
Escaping Quantifiers

• Each quantifier is a metacharacter, and therefore
  they must be escaped if you'd like to explicitly
  match them
• /c\/ matches c, c, c, ... (but not c)
• /get out\??/ matches get out and get out?
• /\bold\/ matches bold, bold, bold, oh
  you get the picture...

11
Metacharacters Grouping Patterns

• Parenthesis ( ( ) ) can group items together for
  the sake of quantifying

print "Continue? (yes or no) " chomp(_
ltSTDINgt) if( /y(es)?/ ) check for y or yes
print "Continuing!\n"
12
Metacharacters Alternation

• Alternative is a fancy regex term for 'or'
• The vertical bar ( ) placed between two
  patterns mean match the left hand side or the
  right hand side
• /textpdfhtml/ means match the string text, or
  the string pdf, or the string html
• From our previous example

if( /(Yy)(eseahah)?/ )