Michael Alipio <[email protected]> asked:
> > my $string = 'boy, pig, 123, 123:412adbd, d0g,
> > lajdlf134><<_ lkadsf !234,';
> >
> > if( $string =~ m/,\s*([^,]*),[^,]*$/ ){
> > print "$1\n";
> > }
>
> How could you guys write this so simple? My regexp was twice this long.
Lots of practice? ;-)
> the regexp after \s* tells perl to match anything (0 or more) that is not
> a comma, right? how come it did not match pig? pig is also followed by
> comma, right? so pig should be captured by ([^,]*), right? I guess perl
> really looks for a match starting from the end of the line.
That would be cool, right? No, my RE is says, I want a comma and maybe some
whitespace after it, then I'll capture all of the following stuff that's not a
comma, then I want another comma and possibly some more stuff that's not a
comma right before the end of the line.
Perl's RE engine starts out by looking at ", pig" (capturing "pig") but the
part where I said "no further comma before the end of the line" prevents it
from matching, since there's a comma after 123 but before the end of the
string. For the same reasons it fails to match at 123, 123:... and d0g.
That's why I think that a "classic C" approach using rindex() and substr()
might be faster in your case:
#!/usr/bin/perl -w
use strict;
my $string = 'boy, pig, 123, 123:412adbd, d0g, lajdlf134><<_ lkadsf !234,';
my $rightComma = rindex( $string, ',' ) - 1;
my $leftComma = rindex( $string, ',', $rightComma );
print substr( $string, $leftComma + 2, $rightComma - $leftComma ), "\n";
__END__
HTH,
Thomas
--
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
http://learn.perl.org/
