[xml] DOM vs xmlReader

["Prashant R" <ramapra gmail com>]

I am at a point where I am investigating the usage for xmlReader as I have a requirement of parsing large xml file that cannot be held in memory

which is why I had a few questions about this.

One of the other things that I have in my environment are that I need to use IO based reading ( as compared to file or memory based reading) 

Due to this i was am using the Push parser mechanism .

Below is the pseudo code for my parsing logic using DOM 

Current DOM  pseudo - code for Parsing using DOM , note that I can only read 1024 bytes from the stream as part of one read operation 

IOObject  mStream;

xmlParserCtxtPtr ctxt;

char pchar[1024];

int res = mStream.ReadBytes(pChar , 4);

if(res)

{

xmlParserCtxtPtr ctxt=   xmlCreatePushParserCtxt(NULL,NULL,

(const char *)pChar, res,NULL);

res = 0;

do 

{

res = mStream.ReadBytes(pChar,fileSizeBytes);

if(res)

errCode = xmlParseChunk(ctxt, (const char *)pChar, res, 0);

} while(res>0 && errCode==XML_ERR_OK);

if( errCode == XML_ERR_OK)

{

xmlParseChunk(ctxt, (const char *)pChar, res, 1))

// Do something , set output 

}

xmlFreeParserCtxt(ctxt);

}

Proposed xmlReader code 

IOObject  mStream;

char pchar[1024]; 

xmlTextReaderPtr reader = xmlReaderForIO (<params( callback functions for IO )>);

while(xmlTextReaderRead(reader))

{

// Do something with this node 

}

xmlFreeTextReader(reader);

My question is that , given this scenario , does xmlreader still save me memory ,compared to DOM  ( in terms of storage allocated for parsing xml) as bytes from the stream would need to still 

be cached to decipher node information.