Regex capture group with + repeat

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

Hi Monks,

Would someone kindly explain why this regex is behaving as it does, and by that I mean why it stops after a word with a trailing ',' and not after a word with a trailing \s? It strikes me that this quantified group of words with a '+' repeat, ...( (\w+[.,;\s]?)+ ) /x; should continue to the end of the string and not stop after the first comma. I need to understand how to stop it at that first comma, (it stops when it should not) as well as how to let it run on to some other anchor. What am I missing?

use v5.20;
my $s = "The zip code 94563 is for the city of Orinda, which is in Con
+tra Costa County, California.";
say "$s";

my ($zip, $words) = $s =~ m/ (\d+) \s+ ( (\w+[.,;\s]?)+ ) /x;

say "$zip $words";
say "$1 $2";
[download]

On my macOS machine that snippet yields:

The zip code 94563 is for the city of Orinda, which is in Contra Costa
+ County, California.
94563 is for the city of Orinda,
94563 is for the city of Orinda,
[download]

This is a simple string example, but I am working on multi-line 'here' documents that are much larger but contain similar \d+ and \w+ constructions.

Thanks for your help.

Comment on Regex capture group with + repeat Select or Download Code

Replies are listed 'Best First'.
Re: Regex capture group with + repeat by haukex (Archbishop) on Dec 23, 2019 at 06:47 UTC
why it stops after a word with a trailing ',' `(\w+[.,;\s]?)+` means to match one or more word characters, followed by zero or one `[.,;\s]` character, and the outer `(...)+` means that the next thing must be one or more word characters, and so on. However, in the string `"Orinda, which"`, the `,` character isn't followed by a word character, it's a space. So it isn't really the difference between comma and space that's causing it to stop, it's more than one non-word character (Update: more precisely, a `[.,;\s]` character followed by a non-word character) - e.g. `..` or `;-` have the same effect. I need to understand how to stop it at that first comma, (it stops when it should not) as well as how to let it run on to some other anchor. If you wanted to stop it at the first comma, then you could remove the comma from the `[.,;\s]` group. A common way to say "match until some character", using the example of `;`, is `[^;];` - this assumes that those characters can't be escaped. One test case isn't really enough though, see Re: How to ask better questions using Test::More and sample data. Update 2:* Minor edits for clarity.	[reply] [d/l] [select]
Re: Regex capture group with + repeat by Anonymous Monk on Dec 23, 2019 at 08:10 UTC
Rxrx? use re 'debug';	[reply]