Adam,
can you re-post this question on the open list ... the eXist development
list is for development issues/tasks, etc.
can I ask also that you include a representative xpath expression and xml
(if the xml u provide is not representative) and I will try to give you some
help.
James Fuller
On 11 July 2010 13:39, Adam Hepner <Ada...@be...> wrote:
> Hi
> I have a (rather big - several hundred MB) XML of the following
> structure stored in the db:
> <krm:catalogue>
> <krm:j id="1637192" pID="1637191" teryt="3716704" powiat="2183789"
> wojewodztwo="2183783"
> kategoria="1">
> <krm:k id="1637194">
> <krm:spec id="2205172"/>
> <krm:spec id="2205241"/>
> <krm:rk id="2205699"/>
> <krm:rk id="2205588"/>
> <krm:rk id="2205589"/>
> <krm:rk id="2205358"/>
> <krm:dsc id="2205171"/>
> <krm:dsc id="2205239"/>
> </krm:k>
> <krm:k id="1637193">
> <krm:spec id="2205213"/>
> <krm:rc id="2206812"/>
> <krm:rk id="2205564"/>
> <krm:dsc id="2205211"/>
> <krm:sc id="2206811"/>
> <krm:sc id="2206808"/>
> <krm:sc id="2206804"/>
> <krm:sc id="2206807"/>
> <krm:sc id="2206806"/>
> <krm:sc id="2206805"/>
> <krm:sc id="2206810"/>
> <krm:sc id="2206795"/>
> <krm:sc id="2206798"/>
> <krm:sc id="2206794"/>
> <krm:sc id="2206791"/>
> <krm:sc id="2206793"/>
> <krm:sc id="2206799"/>
> <krm:sc id="2206797"/>
> <krm:sc id="2206802"/>
> <krm:sc id="2206801"/>
> <krm:sc id="2206800"/>
> <krm:sc id="2206792"/>
> <krm:sc id="2206796"/>
> <krm:sc id="2206809"/>
> </krm:k>
> </krm:j>
> </krm:catalogue>
>
> where there are many nodes of type krm:j, which have many nodes of
> type krm:k. In a rather elaborate process, I filter those nodes, and
> receive a certain subtree - which, therefore, still has the same
> organization. And now I need to count occurrances of every @id in
> every given node type (ie. I need to retrieve information that node
> krm:sc with id 2206800 occurs 5 times, and krm:sc with id 2206801
> occurs 20 times). I found out that just performing a count() on a
> nodeset i(XPath expression result) s surprisingly slow, and can take
> my xquery from below 30 seconds to well over 10 minutes. I have
> indices set up, which sped things up a lot in other contexts, but
> still - counting occurences is a big pain in the lower back. I wonder
> - is there any clever exist-only function which could help mi retrieve
> the counts?
>
> --
> Best regards
> Adam Hepner
>
> tel. 509 093 095
> e-mail: ada...@be...
> http://AdamHepner.pl
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by Sprint
> What will you do first with EVO, the first 4G phone?
> Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
> _______________________________________________
> Exist-development mailing list
> Exi...@li...
> https://lists.sourceforge.net/lists/listinfo/exist-development
>
|
