CODEWITHSUNDEEP
regexperl
jake9115
jake9115

Reputation: 4084

How to make a perl regex and references part of the matching string in the regex?

I am trying to make a perl regex that references something it finds in the query on the fly. I've never tried to make a regex like this, and am not even sure it is possible. Here's my current regex:

$sample =~ s/(+|-)[1][a-zA-Z]{1}//g;

This is meant to take my $sample variable as a query, and delete any section of the string that matches something like +1X with nothing. However, the number defines the number of following characters, all of which I want removed.

So, if my string is foo+3xzybar, the regex should return foobar. How can I grab the matching number in the regex and use it to quantify the number of letters to match? Instead of writing explicit cases like this:

$sample =~ s/(+|-)[1][a-zA-Z]{1}//g;
$sample =~ s/(+|-)[2][a-zA-Z]{2}//g;
$sample =~ s/(+|-)[3][a-zA-Z]{3}//g;
$sample =~ s/(+|-)[4][a-zA-Z]{4}//g;

I want to make something like this:

$sample =~ s/(+|-)[(0-9)][a-zA-Z]{($1)}//g; #(0-9) is meant to match and save the digit in $1

Any help is appreciated!

Upvotes: 1

Views: 161

Answers (2)

user2719058
user2719058

Reputation: 2233

When one is willing to use fancy (and experimental) Perl RE features, the (??{ code }) extended pattern works:

$sample =~ s/[+-](\d)(??{ "[a-zA-Z]{$1}" })//g

Upvotes: 1

woolstar
woolstar

Reputation: 5083

I would probably use a match to breakup the string, and then reassemble it a line later:

$sample =~ /(.*?)[+\-]([0-9])([a-zA-Z]+)(.*)/ ;
$sample= $1 . substr($3, $2) . $4 if $2 ;

So you get everything before the +N, then use +N to offset into the letters right beyond, then tack anything else on afterwards. Repeat as necessary.

Upvotes: 2

Related Questions