You can use \W
to negate the \w
character class, but the problem you've got is that \w
doesn't match your non-ascii letters.
So you need to do something like this instead:
#!/usr/bin/env perluse strict;use warnings;use Data::Dumper;my @folder = ('s,c%','c__pp_p','Monday_øå_Tuesday, Wednesday','Monday & Tuesday','Monday_Tuesday___Wednesday');s/[^\p{Alpha}]+/_/g for @folder;print Dumper \@folder;
Outputs:
$VAR1 = ['s_c_','c_pp_p','Monday_øå_Tuesday_Wednesday','Monday_Tuesday','Monday_Tuesday_Wednesday' ];
This uses a unicode property - these are documented in perldoc perluniprop
- but the long and short of it is, \p{Alpha}
is the unicode alphanumeric set, so much like \w
but internationalised.
Although, it does have a trailing _
on the first line. From your description, that seems to be what you wanted. If not, then... it's probably easier to:
s/_$// for @folder;
than make a more complicated pattern.