Join, split and map using perl for creating new attribs - regex

Question

my $str = "<SampleElement oldattribs=\"sa1 sa2 sa3\">";
$str =~ s#<SampleElement[^>]*oldattribs="([^"]*)"#
my $fulcnt=$&;
my $afids=$1;
my #affs = ();
if($afids =~ m/\s+/) {
#affs = split /\s/, $afids;
my $jnafs = join ",", map { $_=~s/[a-z]*//i, } #affs;
($fulcnt." newattribs=\"$jnafs\"");
}
else {
($fulcnt);
}
#eg;
My Output:
<SampleElement oldattribs="sa1 sa2 sa3" newattribs="1,1,1">
Expected Output:
<SampleElement oldattribs="sa1 sa2 sa3" newattribs="1,2,3">
Someone could point out me where I am doing wrong. Thanks in advance.

Answer 1

Where you're going wrong is earlier than you think - you're parsing XML using regular expressions. XML is contextual, and regex isn't, so it's NEVER going to be better than a dirty hack.
#!/usr/bin/env perl
use strict;
use warnings;
use XML::Twig;
my $twig = XML::Twig -> parse ( \*DATA );
my $sample_elt = $twig -> get_xpath('//SampleElement',0);
my #old_att = split ( ' ', $sample_elt -> att('oldattribs') );
$sample_elt -> set_att('newattribs', join " ", map { /(\d+)/ } #old_att);
$twig -> set_pretty_print ( 'indented_a' );
$twig -> print;
__DATA__
<XML>
<SampleElement oldattribs="sa1 sa2 sa3">
</SampleElement>
</XML>
But to answer the core of your problem - you're misusing map as an iterator here.
map { $_=~s/[a-z]*//i, } #affs;
Because what that is doing is iterating all the elements in #affs and modifying those... but map is just returning the result of the expression - which is 1 because it worked.
If you want to change #affs you'd:
s/[a-z]*//i for #affs;
But if you didn't want to, then the easy answer is to use the r regex flag:
map { s/[a-z]*//ir } #affs;
Or as I've done in my example:
map { /(\d+)/ } #affs;
Which regex matches and captures the numeric part of the string, but as a result the 'captured' text is what's returned.

Answer 2

Here is a simple way to build shown output from the input $str.
Note: The input is in single quotes, not double. Then the \" isn't a problem in the regex.
my $str = '<SampleElement oldattribs=\"sa1 sa2 sa3\">';
# Pull 'sa1 sa2 sa3' string out of it
my ($attrs) = $str =~ /=\\"([^\\]+)/; # " # (turn off bad syntax highlight)
# Build '1,2,3' string from it
my $indices = join ',', map { /(\d+)/ } split ' ', $attrs;
# Extract content between < > so to add to it, put it back together
my ($content) = $str =~ /<(.*)>/;
my $outout = '<' . $content . " newattribs=\"$indices\"" . '>';
This gives the required output.
Some of these can be combined into single statements, if you are into that. For example
my $indices =
join ',', map { /(\d+)/ } split ' ', ($str =~ /"([^\\]+)/)[0]; # "
$str =~ s/<(.*)>/<$1 newattribs=\"$indices\">/;
All of this can be rolled into one regex, but it becomes just unwieldy and hard to maintain.
Above all – this appears to be XML or such ... please don't do it by hand, unless there is literally just a snippet or two. There are excellent parsers.

Answer 3

Found solution on this by searching map function:
my $str = "<SampleElement oldattribs=\"sa1 sa2 sa3\">";
$str=~s#<SampleElement[^>]*oldattribs="([^"]*)"#my $fulcnt=$&; my $afids=$1;
my #affs = ();
if($afids=~m/\s+/)
{
#affs = split /\s/, $afids;
my #newas = join ",", map { (my $foo = $_) =~ s/[a-z]*//i; $foo; } #affs ;
($fulcnt." newattribs=\"#newas\"");
}
else
{
($fulcnt);
}
#eg;
I have updated the below line on my code:
my #newas = join ",", map { (my $foo = $_) =~ s/[a-z]*//i; $foo; } #affs ;
Instead of
my $jnafs = join ",", map { $_=~s/[a-z]*//i, } #affs;
Its working thanks for all.