[xquery-talk] How to optimize XQuery?

Michael Kay mhk at mhk.me.uk
Thu Sep 23 18:17:16 PDT 2004


I notice that you have downloaded a copy of Saxon, let us know how you get
on with it.

Performance questions can only really be answered (a) with knowledge of the
processor, (b) with full details of your query, and (c) with information
that's rather more precise than "a long time" and "big files".

One useful thing to establish is the relationship between the execution time
and the file size: does it double when the file size doubles?

Unlike XSLT, XQuery doesn't define any mechanism within the language for
creating or exploiting indexes, so in general a FLWR expression with a where
clause is likely to involve a serial search. Systems that store the data in
a database and allow you to build indexes are likely to do better than that,
provided you have created the right indexes.

Michael Kay
http://www.saxonica.com/  

> -----Original Message-----
> From: talk-bounces at xquery.com 
> [mailto:talk-bounces at xquery.com] On Behalf Of Robson
> Sent: 21 September 2004 15:23
> To: talk at xquery.com
> Subject: [xquery-talk] How to optimize XQuery?
> 
> Hi all,
> 
> I'm using the xquery implemantion available in 
> http://www.oracle.com/technology/sample_code/tech/xml/xmldb/xm
> ldb_xquerydownload.html. 
> It
> seems to work well, except because the ResultSet take a long time to 
> execute the "resultset.next()", mainly with big files and i'm using 
> "where clauses". Does anyone know how to optimize query or if 
> there is 
> another faster implementation of xquery available to download?
> 
> Thanks in advance,
> cheers,
> 
> Robson Hugo Araújo dos Santos
> 
> 
> 
> _______________________________________________
> talk at xquery.com
> http://xquery.com/mailman/listinfo/talk
> 




More information about the talk mailing list