mnooning has asked for the wisdom of the Perl Monks concerning the following question:

In this narrowed down regex case, each line produces an ERR.

I cannot spot the error. Can you?

Just to be clear, the dollar sign could be one or more other special chars, not just a single dollar sign

my $line1 = 'I:\$AVG';            # Should NOT be okay.
my $line2 = 'I:\$AVG\hello.log';  # Should NOT be okay.

my $skip = '$AVG';
    #------------------------------------------------
    if ($line1 =~ m!\b\Q$skip\E\b!i) {
        print __LINE__." Line 1 Okay!\n";
    } else {
        print __LINE__." Line 1 ERR! $line1 did not match  $skip\n";
    }
    if ($line2 =~ m!\b\Q$skip\E\b!i) {
        print __LINE__." Line 2 Okay!\n";
    } else {
        print __LINE__." Line 2 ERR! $line2 did not match  $skip\n";
    }

__END__
14 Line 1 ERR! I:\$AVG did not match  $AVG
19 Line 2 ERR! I:\$AVG\hello.log did not match  $AVG

[download]

Replies are listed 'Best First'.
Re: regex not matching special char
by kennethk (Abbot) on Dec 13, 2012 at 15:56 UTC
    I think part of your issue is your understanding of what \b means. From perlre:
    A word boundary (\b ) is a spot between two characters that has a \w on one side of it and a \W on the other side of it (in either order), counting the imaginary characters off the beginning and end of the string as matching a \W .

    This means \$AVG will never match /\b\$AVG/ because there is no word boundary between a backslash (\W) and a dollar sign (\W).

    #11929 First ask yourself `How would I do this without a computer?' Then have the computer do it the same way.

[reply]
[d/l]
[select]

      Interesting. A backslash is a word boundary, but a backslashed dollar sign is itself a non-word char itself, and hence is also a word boundary to the string "AVG" which follows it. Rats!

      I need to distinguish between strings such as "$AVG", A$AVG", "A$AVGA". Hence my attempt to do it using \b$AVG\b.

      Am I asking too much of Perl regex?

[reply]

        In addition to looking at the documentation linked by kennethk (and also at perlretut; see in particular the section titled 'Looking ahead and looking behind'), perhaps some insight as to the effect of the  \b (and \B) zero-width word (and non-word) boundary assertions can be gained by split-ting one of the OPed example strings on each assertion: 

        >perl -wMstrict -le
"my $line2 = 'I:\$AVG\hello.log';
 printf qq{'$_'  } for split /\b/, $line2;
 print qq{\n};
 printf qq{'$_'  } for split /\B/, $line2;
"
'I'  ':\$'  'AVG'  '\'  'hello'  '.'  'log'

'I:'  '\'  '$A'  'V'  'G\h'  'e'  'l'  'l'  'o.l'  'o'  'g'

        [download]
[reply]
[d/l]
[select]

        Am I asking too much of Perl regex?

        ;) No, is Perl regex asking too much by asking you to know what you're asking for ? *zing*

        perlrequick

        Luckily YAPE::Regex::Explain handles these 

        use YAPE::Regex::Explain;
print YAPE::Regex::Explain->new(
'\b\$AVG'
)->explain;
__END__
The regular expression:

(?-imsx:\b\$AVG)

matches as follows:

NODE                     EXPLANATION
----------------------------------------------------------------------
(?-imsx:                 group, but do not capture (case-sensitive)
                         (with ^ and $ matching normally) (with . not
                         matching \n) (matching whitespace and #
                         normally):
----------------------------------------------------------------------
  \b                       the boundary between a word char (\w) and
                           something that is not a word char
----------------------------------------------------------------------
  \$                       '$'
----------------------------------------------------------------------
  AVG                      'AVG'
----------------------------------------------------------------------
)                        end of grouping
----------------------------------------------------------------------

        [download]
        use YAPE::Regex::Explain;
print YAPE::Regex::Explain->new(
'\w?\$AVG\b'
)->explain;
__END__The regular expression:

(?-imsx:\w?\$AVG\b)

matches as follows:
  
NODE                     EXPLANATION
----------------------------------------------------------------------
(?-imsx:                 group, but do not capture (case-sensitive)
                         (with ^ and $ matching normally) (with . not
                         matching \n) (matching whitespace and #
                         normally):
----------------------------------------------------------------------
  \w?                      word characters (a-z, A-Z, 0-9, _)
                           (optional (matching the most amount
                           possible))
----------------------------------------------------------------------
  \$                       '$'
----------------------------------------------------------------------
  AVG                      'AVG'
----------------------------------------------------------------------
  \b                       the boundary between a word char (\w) and
                           something that is not a word char
----------------------------------------------------------------------
)                        end of grouping
----------------------------------------------------------------------

        [download]
[reply]
[d/l]
[select]

        What if you change \b to \w?

[reply]
[d/l]
[select]

      For completeness, the code below was the final solution. This solution circumvents the problematic nuance that kennethk explained. 

      if ($line5 =~ m!(\\|\b)\Q$your_item\E(\\|\b)!i) { ... }

      [download]

      Thanks again

[reply]
[d/l]

Back to Seekers of Perl Wisdom