match multiple patterns and extract subpatterns into a array in perl - perl

Question

I have the following string in $str:
assign (rregbus_z_partially_resident | regbus_s_partially_resident | reg_two | )regbus_;
I want to parse this line and only capture all the string that starts with non-word character followed by either reg_\w+ or regbus_\w+ into an array.
so in the above example, i want to capture only
regbus_s_partially_resident and reg_two into a array.
I tried this and it didnot work:
my (#all_matches) = ($str =~ m/\W(reg_\w+)|\W(regbus_\w+)/g);
Since i am trying to use \W, its copying the non-word character also into the array list, which i donot want.

Answer 1

its copying the non-word character also into the array list
No, it doesn't.
$ perl -le'
my $str = "assign (rregbus_z_partially_resident | regbus_s_partially_resident | reg_two | )regbus_;";
my (#all_matches) = ($str =~ m/\W(reg_\w+)|\W(regbus_\w+)/g);
print $_ // "[undef]" for #all_matches;
'
[undef]
regbus_s_partially_resident
reg_two
[undef]
But you do have a problem: You have two captures, so you will get two values per match.
Fix:
my #all_matches;
push #all_matches, $1 // $2 while $str =~ m/\W(reg_\w+)|\W(regbus_\w+)/g;
Far better:
my #all_matches = $str =~ m/\W(reg(?:bus)?_\w+)/g;
Ever better yet:
my #all_matches = $str =~ m/\b(reg(?:bus)?_\w+)/g;

Answer 2

Need a little tweak to your regex
my #all_matches = $str =~ m/\W(reg_\w+|regbus_\w+)/g;
or
my #all_matches = $str =~ m/\W( (?:reg|regbus)_\w+ )/gx;
or even something along the lines of
my #all_matches = $str =~ m/\W( reg(?:bus)?_\w+ )/gx;
The most suitable form depends on what patterns you may need and how this is used.
Or, reduce the regex use to the heart of the problem
my #matches = grep { /^(?:reg_\w+|regbus_\w+)/ } split /\W/, $str;
what may be helpful if your strings and/or requirements grow more complex.