search and replace multiple xml files in a folder using perl

I'm impressed as a newcomer to Perl, you've latched on to map. map is designed to turn an array into a hash - and it can do this by evaluating a code block.

However that's pretty nasty, because it creates code that's hard to follow. Why not instead use a for (or foreach) loop? The key warning sign is 'am I assigning the result of map to a hash (or hashref)?' If the answer is no, then chances are this isn't a good way to do it.

Also: I tend to prefer glob over opendir for this style of iteration operation.

But most importantly of all:

Don't use regular expressions and line based parsing for XML

Please, please, please use an XML Parser to parse XML. Doing so via regular expressions is just nasty - it makes brittle unreliable code. There's a bunch of things in the XML spec that makes semantically identical XML (and therefore 'valid' from a perspective of an upstream system) not match your regular expressions. Things like unary tags, line wrapping and splitting tags across lines.

As an example:

<XML
><some_tag
att1="1"
att2="2"
att3="3"
></some_tag></XML>

Or:

<XML><some_tag att1="1" att2="2" att3="3"></some_tag></XML>

Or:

<XML>
  <some_tag
      att1="1"
      att2="2"
      att3="3"></some_tag>
</XML>

Or:

<XML>
  <some_tag att1="1" att2="2" att3="3"></some_tag>
</XML>

Or:

<XML>
  <some_tag att1="1" att2="2" att3="3"/>
</XML>

All 'say' basically the same thing (technically there's a minor difference between a 'no text' and a 'null text' in the last example), but as I hope you can clearly see - a line and regex based test to encompass all of them would be difficult. Which is why I keep on suggesting - "use a parser" every time this comes up.

With that in mind - you probably don't actually need to remove comments at all - because they're part of the XML spec, and it's far better to handle them as part of the parse process.

I like XML::Twig and perl for this. Other modules exist though, and it may be that you get on with others (like XML::LibXML) instead.

Oh, and there's an error in your XML the line should be:

<?xml version="1.0"?>

Anyway, with that in mind - to answer your question as asked:

Removing comments from some XML

#!/usr/local/bin/perl
use strict;
use warnings;

use XML::Twig;

foreach my $file ( glob("$dir/*.xml") ) {
    my $twig =
        XML::Twig->new( comments => 'drop', pretty_print => 'indented_a' );
    $twig->parsefile($file);
    open( my $output, ">", $file . ".new" ) or warn $!;
    print {$output} $twig->sprint;
    close($output);
}

This will turn your sample XML into:

<?xml version="1.0"?>
<root>
  <subtag>
    <element>This is 1.xml file</element>
  </subtag>
</root>

Deleting 'non comment' elements

If you were wanting to delete something other than comments - bearing in mind that comments are a special case - and instead wanted to say, get rid of a particular element:

XML::Twig->new( pretty_print => 'indented_a',
                twig_handlers => { 'element' => sub { $_ -> delete } } );

Note - this will delete every element tag - you can apply more selective criteria either via an xpath expression (e.g. 'subtag/element') or use a proper subroutine to handle and parse:

sub delete_element_with_file {
    my ( $twig, $element ) = @_;
    if ( $element->text =~ m/file/ ) { $element->delete }
}


my $twig = XML::Twig->new(
    pretty_print  => 'indented_a',
    twig_handlers => { 'subtag/element' => \&delete_element_with_file }
);

##etc.