Thank you for providing more information. Unfortunately it's still a bit unclear to me, for example where do you get the "page4" from, and how does "ind2a" get changed to "ind1"? A few more lines of input might help, or even better, if you could reduce the input down to a small (say 5-20 lines) example that preserves the structure of the input XML, it would help greatly.

At the moment, I can only give the general advice of looking into one of the XML parsing modules, such as XML::LibXML, especially the findnodes function (see XML::LibXML::Node), or perhaps XML::Twig.

(In the future, please indicate in your posting when you update/change it, see How do I change/delete my post?)