Hi Scott,

thanks for your fast reply.
Adding incremental does help so that output is generated
from the beginning.

But the main problem still exist:
    All nodes seems to be kept in memory when closing the transformer.

I guess what's getting in here is DTM, right?
Does DTM has any code to "forget" Nodes after they were parsed?
My XML-Structure is very simple of the form:

<outer>
<typea>
--some-xml-structure-XY-of-about-100-Nodes---
</typea>
<typea>
--some-xml-structure-XY-of-about-100-Nodes---
</typea>
<typea>
--some-xml-structure-XY-of-about-100-Nodes---
</typea>
<typea>
--some-xml-structure-XY-of-about-100-Nodes---
</typea>
...
</outer>

And after parsing a complete typea-Tag I will never need anything from it again.
I guess I will meet the  limit of nodes per document very fast :-( at about 12.000 
Objects
which isn't much when I want to deal with 100.000 Objects.
I guess the limit is because of the way the integer handles are used/generated?

Is it possible to pass "hints" via sax events or via xsl template to "forget" certain 
parts
of the tree? Possibly generating a runtime error if accessing them later on?

The idea was very appealing to me:

        Just generate my stream of XML-Objects using sax-Events and inserting
        simple xslt-Transformations to generate all I need from custom XML to comma
        separated valued :-(
        No need to program java to get the special job done :-(



[EMAIL PROTECTED] wrote:

> Carsten, please see http://xml.apache.org/xalan-j/dtm.html#settings.  You
> want to set "http://xml.apache.org/xalan/features/incremental";  to true.
>
> Be aware that Xalan has a limit of 1,048,575 nodes per document right now.
> Since you have big documents, I'm curious if your getting close to that.
>
> (This is obviously something we going to have to address real soon.)
>
> -scott
>
>
>                     Carsten Heyl
>                     <cal@pixelbox        To:     [EMAIL PROTECTED], Carsten 
>Heyl <[EMAIL PROTECTED]>
>                     x.de>                cc:     (bcc: Scott Boag/CAM/Lotus)
>                     Sent by:             Subject:     XALAN and big files
>                     [EMAIL PROTECTED]
>
>
>                     07/10/2001
>                     04:09 PM
>                     Please
>                     respond to
>                     xalan-dev
>
>
>
> Hi Folks!
>
> I'm impressed by the improvements xalan made since I last tokk a lokk at
> it! Wow!
>
> I have to solve a problem someone may have a hint where to look.
> An instance in my program is generating megabytes of data by pushing the
> data
> into a contenthandler using sax events.
> It's a list of repeating xml-parts having only different attributes.
> Now I wanted to use current xalan as a filter to do some formatting and
> producing a text representation.
> I used the factory to create a transformer and set its output to a file
> StreamResult().
>
> My problem: The transformer seems to hold the complete "Tree" in memory
> before writing anything to the
> file :-(
>
> Any hint where to add a parameter or where to look?
>
> Ciao,
>      Carsten Heyl
>
> --
>   Carsten Heyl                          [EMAIL PROTECTED]
>   Pixelboxx GmbH                       http://www.pixelboxx.de/
>   Ostenhellweg 56-58                    Tel.:   +49 231 53 46 30
> D-44135 Dortmund                        Mobile: +49 174 31 83 608

--
  Carsten Heyl                          [EMAIL PROTECTED]
  Pixelboxx GmbH                        http://www.pixelboxx.de/
  Ostenhellweg 56-58                    Tel.:   +49 231 53 46 30
D-44135 Dortmund                        Mobile: +49 174 31 83 608


Reply via email to