Perl Training

1
Perl Training

• Pattern Matching and Regular Expressions

http//perlwizard.org/perl/week8
2
Agenda

• Homework Review
• Regular Expressions
• Simple True/False Searches
• Getting Shorter
• Those Pesky Backslashes
• Inexact matches
• Characters with Class
• Range Shortcuts
• Special Locations
• Multiplicity
• Match Return values
• Greedy Matching
• Scalar and List Context of m//
• Modifiers
• Global searches
• Grep
• Simple Substitutions
• Using Match Results/Expressions in Second
  Argument
• tr///

3
Homework Problem 1

• Create a program that will open the /www/
  directory on your VServer. Walk through the
  directory structure, making sure that all the
  directories are 755 and files are 644.

View Devens Solution at http//www.netexplorer.or
g/dirwalker.txt
4
Homework Problem 2

• Create a program that will run a Unix command and
  do something with the output (use Pipes)

!/usr/local/bin/perl open(WHO, "who
") while(ltWHOgt) print
5
Homework Problem 3

• Read a list of files from a text file, and print
  the permissions, file size, whether they are
  readable/writable and other usefull file
  information for each file

sub parse_date my _at_time_array ()
local (time) _at__ _at_time_array
localtime(time) Increment Month so
1Jan, 2Feb time_array4
Increment Year, so it's 2000
instead of 100 time_array5 1900
my date_string "time_array4/"
date_string . "time_array3/"
date_string . "time_array5 "
date_string . sprintf("02d",
time_array2) date_string .
sprintf("02d", time_array1)
date_string . sprintf("02d", time_array0)
return date_string
!/usr/local/bin/perl open(FILE,
"/tmp/files.txt") print "Filename\t\tPermissions\
tSize\tLast_Mod\t\tLast_Acc\n" while(ltFILEgt)
filename _ chomp(filename)
my perms (stat (filename))2 07777
my size (stat (filename))7
my last_mod (stat (filename))9
last_mod parse_date(last_mod) my
last_acc (stat (filename))8
last_acc parse_date(last_acc) print
"filename\tperms\t\tsize\tlast_mod\tlast_acc\
n"
6
Regular Expressions

• _at_fields split(/\t/, TheRecord)
• This is the simplest form of Regular Expressions,
  just the exact characters you want to specify,
  such as a single tab in the above example.
• _at_fields split(//, TheRecord) Split each
  field by abccdefff
• _at_fields split(/999/, TheRecord) Split each
  field by 999 abc999cde999fff

7
Simple True-False Searches

• Format
• m// Returns true or false, based on if it finds
  text matching its RegEx
• if(STRING m/REGEX/)
• Example
• if(abcdef m//) print True else
  print False
• Test abcdef
• if(Test m//) print True else print
  False
• if(Test ! m//) If Test does not contain a
  
• find
• if(Test m/find/) Its ok to use variables
• Make SURE that find is not the null string
• If it is, Perl uses whatever is in _, which
  is BAAAAD.

8
Getting Shorter

• You can use other delimiters
• if(Test m)
• if(Test m!!)
• Ill use m// because most people do.
• If you use, the m// form, you dont have to
  include the m.
• if(Test //)
• This is considered the short form and many
  programmers use it.
• To get shorter still, if you are searching in
  Perls special variable _, you dont need the
  at all
• while(ltINFILEgt)
• if(/From/) The line has From in it

9
Those pesky backslashes
How do you search for a / in a
string? if(line /\//) Search for a
vertical bar? if(line /\/) Use Quotemeta
(Requires Perl 5) function print
quotemeta("This string? has ( ) meta chars in
it") prints This\ string\?\ has\ \\ \(\ \)\
meta\ chars\ in\ it Must be escaped
? . ( ) \ (/ is also there
if you use them as your delimiter)
10
Inexact Matches in RegEx

• Use a for or
• if(Test /(xX)/) If Test has an x or an
  X in it
• if(Test /can(dledycer)/) Matches on
  candle, candy or cancer
• Matching any character
• if(Test /N.T/) Matches the letter Nltany
  chargtT
• NET, NeT, NT Not NesT (2 chars)
• Does not work across \n, so it would fail
  if
• Test was N\nT

11
Characters with Class

• if(Test /(xX)/) is the same as if(Test
  /xX/)
• Using will match any characters within the
  brackets
• if(Test /ABCDE/)
• Using in the pattern negates it, so this would
  match any character not in that range
• Which is better?
• if(Test /(0123456789)/)
• if(Test /0123456789/)
• if(Test /0-9/)
• Test for any upper or lowercase letter
• if(Test /A-Za-z/)

12
Range Shortcuts

• Code Replaces Description
• \d 0-9 Any digit
• \w a-zA-Z_0-9 Any alphanumeric
• \s \t\n\r\f Any whitespace character
• \D 0-9 Any non-digit
• \W a-zA-Z_0-9 Any non-alphanumeric
• \S \t\n\r\f Any non whitespace
• if(Test /\d/) Search for a digit
• if(Test /A-Z\d\d/) Search for a capital
  letter followed by two digits
• if(Test /\s\w\w\w\s/) Searches for
  whitespace, followed by 3 letters
• followed by whitespace (a 3-letter word)

13
Special Locations

• indicates the beginning of the string,
  indicates the end of the string
• if(Test /Beg/) If the string starts with
  Beg
• Test Begin here True
• Test I Beg your Pardon False
• if(Test /don/) If the string ends in don
• if(Test /\bY/) If a word starts with a Y
• \b is for word boundary
• if(Test /J\b/) If a word ends with J
• \B is the opposite, i.e. the middle of a
  word

14
Multiplicity

• How do you find one or more hash characters?
• if(Temp //) Not
  general enough
• Symbol Meaning
• Match 1 or more times
• Match 0 or more times
• ? Match 0 or 1 time
• n Match exactly n times
• n, Match at least n times
• n,m Match at least n but not more than m
  times
• (n and m values must be less than 65,536!!)

15
Multiplicity Examples

• if(Test //) Matches one or more
• if(Test /\d3/) Match exactly three digits
• if(Test /\w\d/) Search for one or more
  alphanumeric
• characters followed by one digit
• if(Test /N.T/) NeT, NT, NeT
• if(Test /N.T/) NeT, NesT, NeesT
• if(Test /N.T/) NT, NeT, NesT, NeesT

16
What Perl returns from a match

• Test NESTING
• if(Test /(N.T)/) print True
• print 1\n Prints NEST
• 1 matches the result of the first set of
  parentheses
• 2 matches the result of the second set of
  parentheses
• if(Test /(N.)S(.I)/) 1 is NE, 2 is TI

17
Greedy Matching

• Test NITWITS
• if(Test /(N.T)/) 1 is NITWIT
• Perl uses Greedy matching by default, i.e. match
  as much as possible on the first try, you can
  supress it with the ? after the quantifier
• if(Test /(N.?T)/) 1 is NIT

18
Scalar and List context of m//

• In scalar context, the m// operator returns true
  or false. In list context, it returns the list
  of items returned, such as (1, 2, 3...)
• Test NESTING
• _at_Birds (Test /(N.)S(.I)/) Birds0
  NE, Birds1 TI

19
Modifiers

• Modifier Description
• g Returns each occurrence (global search)
• i Ignores case
• m Allows multiple lines in the string
• o Compiles pattern once (rarely used)
• s Treats as single line
• x Allows whitespace for comments

20
Modifier Examples

• if(Test /(N.T)/i) Matches NesT, Nest,
  nest, net, neT
• if(Test /
• (N.) (? Start with N and any letter)
• S (? ... followed by the letter S)
• (.I) (? ... followed by another letter, then I)
• /x) Same as if(Test /(N.)S(.I)/)
• Its usually more readable to use a regular perl
  comment on the line before your if statement.

21
Global Searches

• Test NESTING
• _at_Birds (Test /N./g)
• Birds0 is NE
• Birds1 is NG
• Handy way to find all the matching strings, if
  you dont know how many there will be.
• i0
• Remembers where in the string it last found
  the last match and starts from there for
  the next search
• while(Test /N./g) i
• print There were i matches\n

22
Grep (A Unix Geek Favorite)

• grep(RegEx, list)
• In List context, returns items that the RegEx is
  true
• In scalar context, returns the number of elements
  that return true
• _at_JustSNames grep(/S/, _at_LastNames) Get all
  names that start with S
• TotSNames grep(/S/, _at_LastNames) Get the
  number of S Names

23
Simple Substitutions

• s/// function format
• s/String to search for/String to replace with/
• Takes the same modifiers as m//
• Count (Test s/e/E/g) All es in Test
  are now E
• Count is number of es changed
• count (Test s/aeiouAEIOU/V/g)
• Change all vowels to V
• If you dont use the , it assumes you want _
• while(ltINFILEgt)
• s/\t//g Change the tabs to vertical bars
• Safer to assign _ to a variable and work
  with the variable
• You can use variables in either part of the s///
  operator
• Test s/tmp/tmp2/g

24
Using Match Results/Expressions in Second
Argument

• Test s/(\d)/Line 1/ Change all lines
  that start with a
• number and change time to Line
• Test s/(,), (.)/2 1/ Change Noble,
  Jason to Jason Noble
• s/// also has the e modifier, which means
  evaluate the second part as an expression rather
  than a string
• Test s/(\d)/1 1/e Turns 23 into 24,
  for all numbers

25
Another look at splitting

• The first argument is a regular expression, not
  just a string.
• Example (from DNS files)
• _at_fields split(/\t\t\t/, TheRecord) What if
  it has 2 or 4 tabs?
• _at_fields split(/\t2,4, TheRecord) Split
  on 2-4 tabs

26
Letter for Letter Translations

• tr/// function translates lots of letters at once
• Test tr/A-Z/a-z/ Translate all uppercase
  to lowercase
• tr/// returns the number of changes it makes.
• Modifiers
• c Complement the first argument (same as in
  m// and s///)
• d Deletes matching characters that are not
  replaced
• s Removes duplicate replace characters
• Test tr/0-9/\n/c Change all non-digits
  into line feeds
• Test tr/0-9//cd Change non-digits into
  nothing
• Test tr/,/,/s Change two or more commas
  into one comma
• y/// is equivalent to tr///

27
Homework

• Im going to take next week off (from teaching
  this class). Instead, each of you needs to
  research a topic in Perl that we either have not
  covered in class, or that you feel needs more
  coverage.
• Perl University class next Monday/Tuesday
• www.cpan.org
• www.perl.com
• www.tpj.com
• Please send me a Power Point Slide presentation
  so that I can put it up on the Perl Website

28
Questions?