I'm trying to write a perl script that removes whitespace from XML tags, but leaves whitespace inside of the values. For example, let's say I have:
<Example>This is an example.</Exampl e>
What I'm looking to accomplish is to knock off the whitespace specifically in </Exampl e>. Since this will be working on an entire XML document, I figured I'd do something with the substitution operator, but I can't quite figure out how to only match whitespace that might be inside of the XML tags themselves.
Any help is greatly appreciated!
Edit: I've added a real example of what is occurring:
not well-formed (invalid token) at line 42, column 25, byte 1456:
<Artist>Eminem</Artist>
<FileName>eminem feat lil wayne - no love -
hotnewhiphop com(2).mp3</ FileName>
========================^
<FileSize>4804478</FileSize>
s!(</?\w+)\s+(\w+\s+/?>)!$1$2!g;
If you want to actually leave whitespace in a tag with attributes, it gets more complex, because whitespace is a legitimate character in a tag. You pretty much have to find the "words" with no equals or space + equals after them and marry them to the previous--unquoted--word.
sub marry_inner_splits {
my $_ = shift;
# fix broken tags
s|^/?(\w+)\s+(\w+)\b(?!\s*=)|$1$2|;
# find the resulting position.
my $pos = index( $_, ' ' );
# return if there is no whitespace.
return $_ if $pos == -1;
# bind the rest of the text to the substring
substr( $_, $pos ) =~ s/(\s*\w+)\s+(\w+\s*=\s*(?:"[^"]+"|'[^']+')\s*)/$1$2/g;
return $_;
}
my $tag_str = q{Some stuff before the tag <ta g attr1="val1" att r2="value #2" /></Escap e>};
$tag_str =~ s/<([^>]+)>/'<' . marry_inner_splits($1) . '>'/ge;
The e flag means that you are*eval*-ing in the replacement part.
I'd strongly recommend avoiding rolling your own XML parsing, or trying to manipulate the XML via regexes etc.
Use one of the many XML parsers available, and save yourself the headaches.
For instance, look at XML::Easy, XML::Smart, XML::TreeBuilder, XML::Twig.
See also http://perl-xml.sourceforge.net/faq/ for a pretty comprehensive FAQ on handling XML with Perl.
