Re: Calculating variable {n} in regex match

[temporal]

in reply to Calculating variable {n} in regex match

It sounds like the OP wants to capture the number of occurrences entirely within the regex, something like:

$foo =~ m/[0-9]({\d+})/
print "$foo has $1 digit(s)";

... if that were possible (it isn't). This would allow the counting match to be thrown in with a longer regex that is matching on other properties as well. This is probably possible through clever use of code evaluation (?{code}) and interpolation within the regex.

However, if all you need to do is count in the regex then you could use the 'goatse' operator:

my $length =()= $foo =~ /[0-9]/g;
print "$foo has $length digit(s)";

Seems better than looping and incrementing, anyway.

[atreyu]

... if that were possible (it isn't). This would allow the counting match to be thrown in with a longer regex that is matching on other properties as well. This is probably possible through clever use of code evaluation (?{code}) and interpolation within the regex.

that's actually kind of hoping for...a clever use of eval - but i couldn't figure that out. anyway, sounds like what i want is not possible, at least in a straight-forward fashion.

[AnomalousMonk]

FWIW:

>perl -wMstrict -le
"my $s = '123 234 3456789 3210 34567';
 ;;
 my @m = $s =~ m{ \b ((\d) (??{ qr{\d{$^N}} })) \b }xmsg;
 print qq{@m};
 ;;
 my $i = 2;
 @m = $s =~ m{ \b (??{ ++$i;  qr{\d{$i}}; }) \b }xmsg;
 print qq{@m};
"
234 2 3210 3
123 3456789

[atreyu]

okay, AnomalousMonk, this works:

# the original number of integers to test (will be one-upped)
my $i = 1; 

# the input string to be tested (a number)
my $foo = shift;

# see if the input string is a two-digit number
my $rc = ($foo =~ m{ \b (??{ ++$i; qr{\d{$i}}; }) \b }xmsg) ?'Y' :'N';

very clever...thanks.

[AnomalousMonk]

But sometimes cleverness can be a curse, so you may want to reserve your thanks until later.