XOM API for Processing XML with Java

[Wolfgang Hoschek <whoschek AT lbl.gov>]

On Jan 19, 2005, at 5:05 PM, shirasu hiroyuki wrote:

> Dear Elliote, Michael and Wolfgang,
>
> Thank you very much for your valuable comments. I learn
> a lot things from them. I must apologize for the delay
> in replying and for my poor English. I write English very
> slowly.

You are most welcome and your work is appreciated very much!

> I know that performance of Saxon for precompiled expressions
> is better than ones of the other engines.
> I planed to check performance of XPath engines for pre-
> compiled XPath expressions first. But I does not understand
> how engines process lifecycle and size of cache for pre-
> compiled expressions.

You can manage those any way you like. The simplest way is to keep the 
compiled query around in your own variables. Or you can use a slightly 
more elaborate cache, typically via some hashmap. One such 
straightforward cache has its lifecycle, thread-safety and management 
options documented at 
http://dsd.lbl.gov/nux/api/nux/xom/pool/XQueryPool.html. A plethora of 
similar schemes could be used, perhaps more elaborate in its eviction 
strategy. But a cache can be made very effectively - I'd be surprised 
if it should turn out to be counter productive in practise...

> To evaluate precompiled expressions
> simply multi times may cause worse performance than non
> precompiled case sometimes (especially, JXPath).
> I would like to study a lot and improve my observation.
>
> Thanks,
> --
> shirasu hiroyuki(hirsh AT s9.dion.ne.jp)
> _______________________________________________
> XOM-interest mailing list
> XOM-interest AT lists.ibiblio.org
> http://lists.ibiblio.org/mailman/listinfo/xom-interest