perl regex: how to capture multi group use only one express?

Question

Text 1 ： 
do not match: nw, n, s and somethingelse.
all directions are: n, w, s and e.
Expect ：{n, w, s, e}


Text 2 ： 
do not match: nw, n, s and somethingelse.
all directions are: nw, sw, se, w, ..., s and e.
Expect ：{nw, sw, se, w, ..., s, e}

Is it possible to capture all directions in one express?

Thaha · Answer 1 · 2015-12-09T08:13:37.433

0

Are you looking for something like this?

my $text = "all directions are: nw, sw, se, w, ..., s and e.";

if( $text =~ /all directions are:\s+(([^,]+,\s+)+)(\w+)\s+and\s+(\w+)\./)
{
    print "$1 $3, $4\n";
}

Output: nw, sw, se, w, ..., s, e

Explanation:

([^,]+,\s+) matches a directory name followed by a comma (,) and some spaces. This can repeat n number of times in the string.

Then we have to match the x and y part. (\w+)\s+and\s+(\w+)\. will take care of that.

edited Dec 09 '15 at 08:13

answered Dec 09 '15 at 07:23

Thaha

303
3
7

Question updated. some conditions missed, it's my bad, sorry. – Jun Zhou Dec 09 '15 at 07:44
Answer Updated. If you just give `all directions are` in the beginning, this will work, right? Or am I missing something here? – Thaha Dec 09 '15 at 08:14

score 0 · Accepted Answer · edited May 23 '17 at 12:07

0

It is not possible capture in one expression as explained here: Python regular expressions - how to capture multiple groups from a wildcard expression?

A soultion to your problem might be this:

use strict;
use warnings;
use Data::Dumper;

my $text = "all directions are: nw, sw, se, w, ..., s and e.";
my @capture;
if($text =~ s/all directions are: (\w+),\s+(.*)/$2/) {
    push @capture, $1;
    while($text =~ s/(\w+),\s+(.*)/$2/) {
        push @capture, $1;
    }
    if($text =~ /(\w+)\s+and\s+(\w+)\./) {
        push @capture, $1;
        push @capture, $2;
    }
}
print Dumper \@capture;

edited May 23 '17 at 12:07

Community

1
1

answered Dec 09 '15 at 07:32

Phyreprooph

517
2
13

Thanks. I can use only one regex in case, it sounds a impossible mission. – Jun Zhou Dec 09 '15 at 07:52
Yeah looks like that's not possible. This is more of a parsing problem. You can't do everything with regex. – Phyreprooph Dec 09 '15 at 07:54

score 0 · Answer 3 · answered Dec 09 '15 at 07:56

While, as Phyreprooph has explained, it's not possible with a single expression, you could use the /g ("global") modifier to match multiple times and yield a list of matches like this:

if(/all directions are:\s*(.*)/) {
    @dir = $1 =~ /(\.\.\.|\b[nsew]{1,2}\b)/g;
    print "{", join(", ", @dir), "}\n";
}

perl regex: how to capture multi group use only one express?

3 Answers3