Re: reading a large XML file using document()

> But this XML file has 36000 lines of data that I have to read. 

do you have evidence that that is actually a problem? I regularly load
files with more lines than that. there may be various ways to make
things faster, use a faster xslt engine, a faster machine or an xml
database for example, but 36000 lines doesn't seem many, and without any
more context hard to suggest anything specific.

unicode.xml available from
http://www.w3.org/2003/entities/2007xml/

has nearly 114000 lines of xml and the xslt transforms in that directory
used to generate documents and entity files, and xslt 2 character maps
for example, just run with saxon from the command line with no special
caching or compilation at all. It's more than fast enough for the job,
but I'm not trying to dynamically regenerate the entire collection every
second (or even every week).


David