[xquery-talk] multiple xml doc reference performance
mike at saxonica.com
Thu May 31 17:07:24 PDT 2007
The performance of a join query like this, and the ways for improving the
performance, depend very much on your choice of XQuery processor. XQuery
doesn't have any equivalent to XSLT's xsl:key which is a hint for creating
indexes; in the absence of this the processor has to work out for itself
what strategy to adopt.
One of the important differences between the two versions of Saxon, Saxon-B
and Saxon-SA, is that Saxon-SA has a more powerful optimizer which is good
at optimizing joins between large datasets. It can make a dramatic
difference for this kind of query.
> -----Original Message-----
> From: talk-bounces at x-query.com
> [mailto:talk-bounces at x-query.com] On Behalf Of Charlie
> Sent: 31 May 2007 15:38
> To: talk at x-query.com
> Subject: [xquery-talk] multiple xml doc reference performance
> I have two xml files, one is about 200M and the second is
> 1.5M. I tried to go through the first big file and check the
> info in the 2nd file for every iteration, so it looks like
> two for loops.
> My question is what the better solutions are to make this
> kind of query faster. I have some other queries running on
> one 400M file within two minutes, but this query using two
> smaller files takes much longer.
> Thanks for any suggestions,
> Fussy? Opinionated? Impossible to please? Perfect. Join
> Yahoo!'s user panel and lay it on us.
> talk at x-query.com
More information about the talk