You should check out vtd-xml, which is ideally suited for the task you 
described...
http://vtd-xml.sf.net



"fuel" <ajaykumarns@g...> wrote in message 
news:0ced59a0-c563-4493-8091-f0f72848bd89@l......
> Hello,
>       I have two big xml files (around 50-60 MB) each. and I need to
> process the data within each of them. The problem is, I need to
> process each node and compare with the other nodes in the other xml
> file. After iterating through all the nodes, I need to find those
> nodes which have changed or which have been newly introduced.
> Assume the following xml structure,
>
> <?xml version="1.0"?>
> <root>
>    <nodeToProcess>
>
>    </nodeToProcess>
>     .....
> </root>
>
> I have two such xml files. I keep one xml file as the reference and
> compare it with the other. To solve this problem, I thought, I could
> use XPath. However, for now, only DOM based XPath processors are
> there. Since the file is very huge, I dont think I can afford DOM.
> ( Memory constraint )
>
> How can I approach this problem ? what would be the right way to start
> with.
>
> P.S ( I am trying to access these elements through Java)
>
>