[xquery-talk] performance figures for xquery processors

Michael Kay mhk at mhk.me.uk
Tue Feb 7 17:17:22 PST 2006

> > Wonder, if any one has done performance testing on
> > xquery processors that have java API and has some data
> > to share with.

I can't add any data - comparative measures are only worth having if done by
third parties, because any vendor can make their own product go faster than
other people's - but I would make one or two observations.

First, take care to separate query compilation time, document building time,
and query execution time. Vendors make different trade-offs between these
factors, because they are optimizing for different scenarios.

Second, make sure that you run things for long enough to eliminate start-up
overheads - Java takes a while to get up to speed.

It's been suggested on this thread that the Saxon tree format takes more
space than for example DTM. I would be surprised if there were a significant
difference for most cases, so I'd be interested to see the evidence.
However, it's good to see people discussing space as well as time: both are
important (and savings in memory often translate into savings in time,
especially when memory is short).

Michael Kay

More information about the talk mailing list