At 2005-07-11 11:43 +0200, Hondros, Constantine wrote:
I am pre-processing batches of about 1000 XML files at a time using Saxon.
Part of the pre-process involves aggregating linked XML documents into the
current document. Naturally, I use the document() function for this:
...
How would you optimise this? Would a deep-copy with <xsl:copy-of> be faster?
Or am I better off writing my own processor for this aggregating step (easy
enough).

Since the XSLT processor is obliged to keep all document() node trees 
around somehow (it doesn't know when you might need a given tree again in 
the transform and any generated identifiers for the nodes need to be 
persistent in the transform), I would recommend a simple SAX process for 
this kind of aggregation.  Not only would it be faster (no building of the 
node tree), it would have a very small footprint (no persistence of any 
input documents).



Some XSLT processors will allow you to pass your output SAX events as input 
to an XSLT transform so in such a case you wouldn't even have the 
aggregated file sitting around to worry about.



I hope this helps.



. . . . . . . . . Ken




--
World-wide on-site corporate, govt. & user group XML/XSL training.
G. Ken Holman                 mailto:gkholman@xxxxxxxxxxxxxxxxxxxx
Crane Softwrights Ltd.          http://www.CraneSoftwrights.com/s/
Box 266, Kars, Ontario CANADA K0A-2E0    +1(613)489-0999 (F:-0995)
Male Breast Cancer Awareness  http://www.CraneSoftwrights.com/s/bc
Legal business disclaimers:  http://www.CraneSoftwrights.com/legal