[
Lists Home |
Date Index |
Thread Index
]
Basically what I want is: once a SAX event is generated, it cannot be
generated again. So either we can cache a whole DOM tree from SAX events,
and query on it, but that is impossible in our scenario.
This approach really be used for all scenarios, rather we need some
scenario restriction, such as given a query, we compute the result against
only the rest of the SAX events, not the prior sax events etc..
So I believe a general purpose XQuery engine will never have these
characteristics.
best, murali.
On Thu, 20 Nov 2003, Oleg Tkachenko wrote:
> Murali Mani wrote:
>
> > I do not think Saxon is what I am looking for. What I want is I give a
> > xpath expression/xquery etc, and some one gives it sax events. I want the
> > query processor to operate on this stream of sax events from this point
> > on, and give me values that satisfy the path expression/query.
>
> So take a look at "Using XPath Expressions" chapter in Saxon
> documentation at
> http://saxon.sourceforge.net/saxon6.5.3/api-guide.html#Expressions
> (or http://saxon.sourceforge.net/saxon7.8/api-guide.html#Expressions for
> Saxon 7.X).
> It sounds exactly what are you looking for.
>
> > it says it can rebuild trees etc, which is not really possible in my
> > scenario..
>
> Well, rebuilding trees usually implies extensive querying, so usually
> any XSLT processor exposes XPath API too.
>
> > think of it more like your data is continuously streaming in,
> > and is "infinitely" large..
> Sounds like you don't want to build in-memory tree? Then XPath might be
> not appropriate query language for you, because it operates on tree
> level, hence allows reverse axes etc.
> --
> Oleg Tkachenko
> http://www.tkachenko.com/blog
> Multiconn Technologies, Israel
>
|
