If speed is critical, you'll find that XML::LibXML is much faster but a bit more 'bleeding edge'.

I'll hazard a guess that XML::Simple is reading the whole file, parsing it, creating a data structure and only then returning. That's always going to take time, and for a large document it'll produce a *huge* data structure which might even make your machine swap - hence the sloooooowness.

You might want to look at a streaming parser instead, which reads the file a bit at a time, generating a series of events for you to handle. It'll be a bit more work but will save an awful lot of memory. Even if it's not any faster over all, it'll at least start giving you data sooner so will *appear* to be faster!

It's hard to give you an advice if we do not know what do you plan to do with the data from the XML and/or how much of the data do you even plan to use!

Apart from the modules others already suggested you might try XML::Rules. (Yeah, it's mine, if I don't advertise it, no one will.) It'll allow you to filter the XML as it's being parsed so then you end up with only the stuff you are interested in instead of a huuuuge, deep tree containing mostly stuff you have no use for and that only occupies the memory and maybe even forces your computer to start swapping memory.

You can think of XML::Rules as XML::Simple on steroids, in XML::Simple you can say that you want these tags to be represented as arrays even if there is just one and to use an attribute as the hash key, but that's about it. XML::Rules will allow you to specify that for this tag you want just the content, for that one just this attribute, that you only want the dat in this tag if the attribute foo's value is 'bar', etc. etc. etc.

SYNOPSIS
           use XML::Simple;

           my $ref = XMLin([<xml file or string>] [, <options>]);
[download]

Sounds like a good idea, if not, XML::Twig can do just that .. see the twig_roots mode ..

It comes from Perl Cookbook, section: 22.8. Processing Files Larger Than Available Memory.

There is a small example of how to use XML::Twig to read in only portions of a large XML file.

Note: the URL above will point to something different tomorrow.