jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Alex Lukin <lu...@stu.cn.ua>
Subject Re: Re: Q. Xpath performance again
Date Thu, 07 Feb 2008 20:31:17 GMT
Hi again! 

First,
I place link to yor explantions on the wiki. http://wiki.apache.org/jackrabbit/Performance
Guys, please do not beat me, it's my first wiki post ever :)

Second,
Your asdvise didnt helped. I removed all test extractors from <SearchIndex> section
and 
placed  <param name="respectDocumentOrder" value="false"/> there.
I re-imported all content in ampty repository with new config and results are the same.

I do not need to check any properties of nodes, I just want get all nodes by some path in
XPath query.
Actually I have exactly 1234 nodes and it takes about minute on my 1.8 GHz notebook to get
them into QueryResult in regular run.
Debugger "sits" on "QueryResult qr = q.execute();" line for ten minutes and processor cooler
simply screams. 

Code:
 
            QueryManager qmgr = rsl.getJCRSession(httpSession).getWorkspace().getQueryManager();
            Query q = qmgr.createQuery(xQuery, queryLang);
            QueryResult qr = q.execute();  //<---this line takes allmost all time --
            NodeIterator ni = qr.getNodes();
            int counter = 0;
            while (ni.hasNext()) {
                Node n = ni.nextNode();
            .....
            }

Query is quite simple: "//uk/news//*/pib:pibdocument"

Nodes of my content are not comlicated but cosists of 2 levels of 2-3 subnodes with 5-10 properties.
Properies are mostly plain text and size is about 1K

I jsut do not know what to do. Such slow behavior breakes all my application.
  
-- 
SY, Alex Lukin
RIPE NIC HDL: LEXA1-RIPE

Mime
View raw message