Using regular expressions

1
Using regular expressions

• Search for a single occurrence of a specific
  string.
• Search for all occurrences of a string.
• Approximate string matching.

2
Forming RegExps

• Strings
• Variables
• Patterns

3
Strings and Variables

• /Joey Ramone/ - match a specific string.
• /name/, where name Joey Ramone - match the
  string stored in a variable.
• /Joey name/ - matching a pattern defined by a
  mixture of strings and variables.

4
Character classes

• abc match abc
• . match any single character (i.e. a.b).
• abc match a or b or c
• 0123456789 match 0 or 1 or or 9
• 0-9 same as previous
• a-z match a or b or or z
• A-Z same as previous only with caps
• match any single occurrence of any of the
  characters found within.
• 0-9a-zA-Z- match any alphanumeric or the
  minus sign

5
Negated character classes

• 0-9 match any single character that is not a
  numeric digit
• aeiouAEIOU match any single character that
  is not a vowel
• Works only for single characters
• Well discuss matching negated strings of
  characters later.

6
Escape characters

• \ - use the backslash to match any special
  character as the character itself.
• /\name/ - match the literal string name.
• /a\.b/ - match the literal string a.b rather
  than a followed by any character, followed by
  b.

7
Convenience character classes

• \d (a digit) - 0-9
• \D (digits, not!) - 0-9
• \w (word char) - a-zA-Z0-9_
• \W (words, not!) - a-zA-Z0-9_
• \s (space char) - \r\t\n\f
• \S (space, not!) - \r\t\n\f

8
Sequences

• - one or more of preceding pattern
• /a-zA-Z/ (match a string of alpha characters
  such as a name).
• ? (match zero or one instance of preceding
  character).
• /a-zA-Z-?a-zA-Z (Now we can match
  hyphenated names).

9
Sequences

• (match zero or more of preceding pattern)
• Example list of names
• George Harrison
• Paul McCartney
• Richard Ringo Starkey
• John Winston Lennon
• /a-zA-Z a-zA-Z/ (match first and last name)
• /a-zA-Z a-zA-Z\ a-zA-Z/ (match first
  name, middle name, if it exists, and last name)

10
Sequences

• k match k instances of preceding pattern.
• Example floating point numbers to 2 decimal
  places
• /0-9\.0-92
• k,j match at least k instances of preceding
  pattern, but no more than j.
• Example floating point numbers that may or may
  not have a decimal component.
• /0-9\.?0-90,2/

11
Grouping

• /(JohnPaulGeorgeRingo)/ matches any one of
  either John, Paul, George, or Ringo
• /((JohnPaulGeorgeRingo) )/
• Matches the Beatles names listed in any order.
• John Paul George Ringo
• Paul George John Ringo
• Ringo Paul George John
• Actually, this will also match
• Paul Paul Paul Paul Paul Paul Paul Paul Paul
• Be careful about what assumptions you make.

12
Problem

• Write a regular expression that will match social
  security number.
• Format 555-55-5555

13
A solution

• /0-93-0-92-0-94/

14
Problem

• Write a regular expression that will match a
  phone number.
• Formats
• 319-337-3663
• 319.337.3663

15
A solution

• /0-93\.-0-93\.-0-94

16
Add another format

• 3193373663

17
A solution

• /0-93\.-?0-93\.-?0-94/

18
Problem

• Write a regular expression that will match an
  email address.
• Legal characters for names are
• Letters, numbers, -, and _
• Legal characters for domain names are
• Letters only
• Assume form username_at_machine.domain.suffix

19
A solution

• /a-z0-9-_\_at_a-z(\.a-z)2/
• More general version
• /a-z0-9-_\_at_a-z(\.a-z)/

20
Problem

• Write a regular expression that will match an
  HTML anchor start tag.
• Assume anchor tag is of the form
• lta hrefsome urlgtsome anchor textlt/agt

21
A solution

• /lta hrefgt/
• Actually, quotes are not required
• So it should be
• /lta href?gt?gt/
• How would we assign the url to a variable?

22
A solution

• (url)
• (htmlText m/lta href?gt?gt/)

23
Take Away

• There is almost always a pattern that will match
  what you want it to match.
• The best way to learn is to simply jump in and
  start writing your own patterns.
• If you have a question about how to construct
  one, feel free to ask me.
• One typically learns Perl by asking people with
  more experience.