[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: an XQuery API for Java

• To: "'Burak Emir'" <Burak.Emir@e...>
• Subject: RE: an XQuery API for Java
• From: "Michael Kay" <michael.h.kay@n...>
• Date: Tue, 6 Jul 2004 12:08:50 +0100
• Cc: <xml-dev@l...>
• In-reply-to: <40EA857E.1080502@e...>
• Thread-index: AcRjR7Cb3mAqLyjNTTizsMoX7GegfwAAQgXg

> 
> Does the use of SequenceIterator imply that the result is 
> materialized 
> in memory, excluding delayed execution ?
> Couldn't there be an implementation of the interface that just defers 
> execution ?

The list underpinning the SequenceIterator isn't materialized, but if there
are nodes in the sequence, then the trees rooted at these nodes are
materialized. It's possible in theory of course to have a Node
implementation that does deferred materialization, but Saxon doesn't do this
currently.

> For document transformations, I think a JAXP result does not 
> help much 
> either - because if I understood it correctly, it can only be 
> used for 
> the "last step" - you cannot run queries or transformations.

Actually, one of the nice features of the JAXP interface (which appears not
to have been studied closely by the XQJ designers) is the way it allows
transformations to be pipelined.

In XSLT of course the question of how a stylesheet is supplied to the
interface has an obvious answer: it is supplied as an XML document, using
any representation of an XML document that one chooses.

Michael Kay

• Follow-Ups:
  • Re: an XQuery API for Java
    • From: Burak Emir <Burak.Emir@e...>

• References:
  • Re: an XQuery API for Java
    • From: Burak Emir <Burak.Emir@e...>

• Prev by Date: Re: an XQuery API for Java
• Next by Date: Software Development 2005 West Call for Papers
• Previous by thread: Re: an XQuery API for Java
• Next by thread: Re: an XQuery API for Java
• Index(es):
  • Date
  • Thread