Regular expression (/<(\w+)\s+(.*?)>/) need improvement

Question

Hi, There is a sub to handle the Type and Value.

sub parse_type_value_specifier { 
    my $tvs = shift; 
    my ($type, $value) = $tvs =~ /<(\w+)\s+(.*?)>/;
    return $type, $value; 
}

It should suit for three formats below.

<B 0> - works, return $type = (B) and $value = (0)
<A[1..80] ""> - doesn't work, need return $type = A[1..80] and $value = () # empty
<A[1..80] "hello"> - doesn't work. need return $type = A[1..80] and $value = (hello)

/<(\w+)\s+(.*?)>/ Thank you.

Answer 1

It sounds like you want to ignore "s. Run it through another regex to strip those out first.

Answer 2

How about

/<([\w\[\].]+)\s*"?([^">]*)"?>/

or /<(\w+)\s*"?([^">]*)"?>/ if your A[1..80] means \w length 1 to 80

Answer 3

Try this:

/<(\w{1,80})\s*(?:\s([^\s">]+|"[^"]*"))?>/

Now if the match of the second grouping starts with a ", remove it from the start and the end and you have the plain value.

Answer 4

Try this

<(.+) +"?(.*?)"?>

Answer 5

Your regex is 99% correct, problem is that \w will not match literal square braces []. just repace \w with a suitable character class [\w\[\]\.]+

<([\w\[\]\.]+)\s+(.*?)>

Answer 6

The following "works" for the input you show but you should provide a more complete spec:

#!/usr/bin/perl

use strict; use warnings;

while ( <DATA> ) {
    if ( my ($type, $value) = /^<([A-Z])(?:\[.+\])?\s+"?(\w*)"?>/ ) {
        print "\$type = $type\t\$value = $value\n";
    }
}

__DATA__
<B 0>
<A[1..80] "">
<A[1..80] "hello">

Output:

$type = B       $value = 0
$type = A       $value =
$type = A       $value = hello