Handling large files

[Mark]

I currently use TinyXML to export and import data from a server application to 
it's client. I load the export file into a TiXmlDocument. The classes in the 
client have a constructor that takes a TiXmlElement* as the argument, and the 
class properties are assigned from that.

That works fine for small documents but with larger documents it can take up a 
lot of memory, so I thought I'd try using SAX with Xerces. I'd still like to 
pass a node object (whatever the Xerces equivalent is) to the client classes but 
I can't see that the events for the SAX parser give me anyway of accessing the 
actual xml, aside from rebuilding the node. Rebuilding the node becomes 
difficult because I need an indication of depth for nested nodes.

Is there any way of parsing a file one node at a time without descending to 
child nodes? Or rebuilding a node's xml?

TIA