[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: Indexing XML documents
On Mon, 03 Jul 2000, Lars Martin wrote:
> On Mon, 03 Jul 2000, you wrote:
> > They are using their own query language. Because of the lack of any
> > alternative we have to use XPath. But XPath is not intended to be used for
> > querying a la SQL. IMO indexing isn't useful for XPath per se. XPath
> > much deals with paths like " /PLAY/ACT/TITLE". Using the DTD is the way to
> > speed up path processing. And for the real searches like
> > "[../SCENE/SPEECH/SPEAKER='CLEOPATRA']" simple indexes are not useful because
> > the result set is most often qualified (read: one cannot simply use the SPEAKER
> > index, which probably contains all SPEAKERS of the document, in the above
> > example)
>
> First thought on this: You don't know any DTD inside the database.
I didn't speak about databases. I speak about XPath performance in general.
Anyway, why do you think the DTD is not known inside the database per se?
>
> > So the the question is, could indexing help to speed up XPath anyhow or is it
> > "inherently non-indexable".
>
> Some weeks ago David Duddleston reported that eXelons XPath queries on indexed
> documents are much more faster than normal XPath queries. So this question is
> answered already!
David reported that eXelons XPath is very fast. But you don't know what they
have done to make it that fast. They may call it indexing because everybody
knows that indexing makes databases fast. But this says nothing about their
actual technology. Or am I wrong? Do you know what actually works inside eXelon?
Falko
--
______________________________________________________________________
Falko Braeutigam mailto:falko@softwarebuero.de
softwarebuero m&b (SMB) http://www.softwarebuero.de