Re: [xsl] Reducing memeory overheads with xsl

["Simon Kelly" <kelly@xxxxxxxxxx>]

I'm using the javax.xml.transform.Transformer() if memory and spelling serve
me well.

I am trying to stay away from having any thing thats not 'standard' in all
systems, so I'm sticking to packages that will 'come as standard' in the Sun
Java development enviroment.

Cheers

Simon


----- Original Message -----
From: "Michael Kay" <mhk@xxxxxxxxx>
To: <xsl-list@xxxxxxxxxxxxxxxxxxxxxx>
Sent: Thursday, April 17, 2003 11:30 AM
Subject: RE: [xsl] Reducing memeory overheads with xsl


> The answer to this is obviously very processor-dependent. Do tell us
> which XSLT processor you are using.
>
> With Saxon, my experience has been that a 25Mb document will transform
> in 512Mb of memory. I've heard of people transforming documents of 100Mb
> and beyond. But I do question whether this is the right design approach
> when things grow to that kind of size.
>
> Michael Kay
> Software AG
> home: Michael.H.Kay@xxxxxxxxxxxx
> work: Michael.Kay@xxxxxxxxxxxxxx
>
> > -----Original Message-----
> > From: owner-xsl-list@xxxxxxxxxxxxxxxxxxxxxx
> > [mailto:owner-xsl-list@xxxxxxxxxxxxxxxxxxxxxx] On Behalf Of
> > Simon Kelly
> > Sent: 17 April 2003 08:29
> > To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
> > Subject: [xsl] Reducing memeory overheads with xsl
> >
> >
> > Hi all,
> >
> > I currently have the problem that my system is running out of
> > memeory when processing my xslt.  I have set the max memory
> > to 526MB (over the 512 I
> > have) in a vain attempt to get the xsl-processor to use every
> > ounce of memory.  But it is still keeling over at the point
> > when I need to work on about 40,000+ tags.
> >
> > I get this structure form the db
> >
> > <rowset>
> >   <row>
> >     <timestamp>1</timestamp>
> >     <sig1>1</sig1>
> >     <sig2>2</sig2>
> >     ....
> >     <sign>n</sign>
> >   </row>
> >   ....
> >   <row>
> >     <timestamp>N</timestamp>
> >     <sig1>1</sig1>
> >     <sig2>2</sig2>
> >     ....
> >     <sign>n</sign>
> >   </row>
> > </rowset>
> >
> > and what I need to end up with is a <measuring sequence>
> > containing one of the following for each row.
> >
> > <measuringdata>
> >   <timestamp>1</timestamp>
> >   <sigval>
> >     <label>sig1</label>
> >     <value>1</value>
> >   </sigval>
> >   ....
> >   <sigval>
> >     <label>sign</label>
> >     <value>n</value>
> >   </sigval>
> > </measuringdata>
> >
> > The whole final file should only be around the 6MB mark, and
> > I'm creating it in 1/2 a gig of ram, so I have a couple of questions.
> >
> > 1)  Why does it run out of memory if the file sizes are 1/64
> > of max memory?
> > 2)  Is there a way to generate the file in a more memory
> > efficient way??
> >
> > Your help, as always, would be most appreciated.
> >
> > Cheers
> >
> > Simon
> >
> >
> > Institut fuer
> > Prozessdatenverarbeitung
> > und Elektronik,
> > Forschungszentrum Karlsruhe GmbH,
> > Postfach 3640,
> > D-76021 Karlsruhe,
> > Germany.
> >
> > Tel: (+49)/7247 82-4042
> > E-mail : kelly@xxxxxxxxxx
> >
> >
> >  XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list
> >
>
>
>  XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list
>
>


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list