From: James Britt <james_b@...>
Date: 2005-06-24T14:10:24+09:00
Subject: Re: REXML Optimization

Bucco wrote:
> Just to help me complete the learning process, I wish to pose the
> question to the group: Is there a better way to do this, and is there
> more optimization I can do to my code?


The DOM is not a database, and it shows.

XPath queries can get real slow as the document size grows.

Suggestion: Read and parse the XML once, and store it internally in a 
format better suited for queries.  XML is great for all sorts of things, 
particularly for inter-app data exchange, but once the data is inside 
your system that value drops.  So, if the code is mainly concerned with 
executing queries and such, slurp in the XML and stash it in some 
optimized internal structure.  Maybe use Madeleine for in-memory storage 
and queries.

If need be, add code to serialize the data back to XML for persistence 
when the app is shut down.

Try to compute the start-up cost of the parsing and restructuring and 
indexing the data right up front, versus the cost of running XPath calls 
over and over.  See if it gains you anything.



James

-- 

http://www.ruby-doc.org - The Ruby Documentation Site
http://www.rubyxml.com  - News, Articles, and Listings for Ruby & XML
http://www.rubystuff.com - The Ruby Store for Ruby Stuff
http://www.jamesbritt.com  - Playing with Better Toys