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

Re: Compressed XML and XSLT

• From: Jonathan Robie <jonathan.robie@r...>
• To: Jeff Lowery <jlowery@m...>
• Date: Mon, 22 Oct 2007 14:42:04 -0400

An XQuery implementation that supports document projection, such as 
DataDirect XQuery, creates only the portion of the input document needed 
to process the query. Document projection was originated by the Galax 
people, and their processor was the first I know of to do document 
projection. If you use streaming output, you can be very efficient in 
your memory usage. I've successfully queried 60 Gig files, not instantly 
but reasonably quickly.

I know Saxon can do XSLT transforms in this range. See 
http://www.saxonica.com/documentation/sourcedocs/serial.html (or wait 
for Mike to chip in ;->)

Jonathan

Jeff Lowery wrote:
> Is there an solution that can take an XML file in a compressed format 
> that will perform the minimal in-memory expansion needed to apply 
> various extraction filters?
>  
> The problem is this: given a large, compressed XML file, I want to 
> apply varous filtering templates to produce an abstract of the 
> original, which may be many times smaller. With that smaller input, I 
> can then apply further transforms that an entire document in-memory.
>  
> The goal is to produce an abstracted and reformatted XML file with 
> minimal disk space and memory usage. It would seem that whether the 
> compressed input would need to be fully expanded iin-memory would 
> depend on both the compression method and queries involved.
>  
> I see there's an XQuery project called  XQueC that might apply here, 
> but I don't know that it ever escaped research stage.
>  
> Regards,
>  
> Jeff Lowery
>  

• References:
  • Compressed XML and XSLT
    • From: "Jeff Lowery" <jlowery@m...>