problem using regexp

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

how do i match a particular alphabet using regexp? I cannot print the second occurrence of 'a' in my string in the following piece of code,as a beginner I cannot find where the problem is and why it is happening?

#! usr/bin/perl
use warnings;
use strict;
my @vars;
$_='1: A a A silly a sentence a (495,a) *BUT* one a which is a will be
+ useful. (3)';
if (@vars= /(\s)(a)(\s)/ig){
    print "the text matches the pattern '\$pattern'.\n";
    foreach (@vars) {
        print "Group: $_\n";
    }
}
else {
print " it was not found. \n";
}
print "@vars\n";
[download]

Comment on problem using regexp Download Code

Replies are listed 'Best First'.
Re: problem using regexp by AnomalousMonk (Archbishop) on Jul 28, 2012 at 01:33 UTC
`if (@vars= /(\s)(a)(\s)/ig){ ... }` [download] Another approach to handling overlapping matches (and, IMHO, better because less error-prone and better integrated into the general regex expression than the manipulation of pos) is the use of a look-ahead assertion to wrap capturing groups: see `"(?=pattern)"` in the Look-Around Assertions sub-section of perlre Extended Patterns. Replacing `/(\s)(a)(\s)/ig` with `/(?=(\s)(a)(\s))/ig` in the code quoted above from the OP produces the desired capture (tested).	[reply] [d/l] [select]
Re^2: problem using regexp by artifact (Novice) on Jul 29, 2012 at 16:15 UTC
thanks for replying so fast ...and also it encouraged me to find different techniques.	[reply]
Re: problem using regexp by choroba (Cardinal) on Jul 27, 2012 at 20:30 UTC
The first match eats up the space. The second 'a' therefore is not matched, because there is no space before it. If you do not want the space to be eaten, you can use a look-around: `/\s(a)(?=\s)/ig` [download]	[reply] [d/l]
Re^2: problem using regexp by artifact (Novice) on Jul 29, 2012 at 16:01 UTC
thanks ,it works!!	[reply]
Re: problem using regexp by Grimy (Pilgrim) on Jul 27, 2012 at 20:36 UTC
The problem here is overlapping matches. The space between the first and the second occurence of /a/i is part of both possible matches. Since it is gobbled by the first match, the second /a/i doesn't match. It is possible to "rewind" your regex by assigning to `pos`. See 'pos' in perldoc. `#! usr/bin/perl use strict; use warnings; $_=' A a A '; while (/(\s)(a)(\s)/ig) { print "The text matches the pattern '$&' at position " . pos() . " +.\n"; # uncomment this to rewind #pos() = pos() - 1; }` [download]	[reply] [d/l] [select]
Re^2: problem using regexp by artifact (Novice) on Jul 29, 2012 at 16:05 UTC
it works!! it also made me learn few other concepts.	[reply]