Re: [xml] GSoC using libxml for indexing XML document

[Liam R E Quin <liam holoweb net>]

On Sun, 2011-06-05 at 11:19 +0800, Daniel Veillard wrote:
> On Fri, Jun 03, 2011 at 05:34:03PM +0200, Tomáš Pospíšil wrote:
> > Hi Daniel and all hackers,
> >
> > I'm GSoC student creating new XML index in PostgreSQL which use
> LibXML for handling XML documents. My idea about index is about to use
> node offsets and Patricia Trie for mapping structural information to
> our internal representation tree index. So how can I get offset of
> nodes?

Before you get too far you might want to look at xqilla and dbxml -
presumably your goal will be to support XQuery, and you'll want the
parser and XPath 2 engine for that (unless PostgreSQL already has XQuery
support?).  You might continue on your current course, of course, but
take a look.  Maybe you did that already.

Patricia trees can be fast (PAT used to use them, years ago, a spin-off
from the University of Waterloo called OpenText), but they are difficult
to update, so if you plan to support XQuery Update, e.g. to replace a
single text node in a 500 MByte XML document, it might not be the best
choice.

Liam


-- 
Liam Quin - XML Activity Lead, W3C, http://www.w3.org/People/Quin/
Pictures from old books: http://fromoldbooks.org/