I did noticed (or so it seems to me) that changing the select
clauses to
*[generat-id(.)=generate-id(key('tags', name())]
will slow down the processing...
I think Jeni Tennison did some comparison a while ago of the costs of the
two different XSLT 1.0 approaches for comparing node identity (the XPath 2.0
"is" operator), and found that some processors were much faster on
count(A|B)=count(A), while others were faster on
generate-id(A)=generate-id(B). So it depends which processor you are using.
(It probably depends on the product version as well. After Jeni's study, I
introduced an optimization into Saxon to recognize
generate-id(A)=generate-id(B) and translate it directly into an identity
comparison without actually generating the IDs.)
Michael Kay
http://www.saxonica.com/
--~------------------------------------------------------------------
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list
To unsubscribe, go to: http://lists.mulberrytech.com/xsl-list/
or e-mail: <mailto:xsl-list-unsubscribe(_at_)lists(_dot_)mulberrytech(_dot_)com>
--~--