jackrabbit-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Christoph Kiehl <christ...@sulu3000.de>
Subject Re: better sql performance? [was: Re: poll on jcr query usage]
Date Mon, 02 Apr 2007 14:16:37 GMT
David Nuescheler wrote:
> hi christoph,
> 
>> I'm mostly using XPath, but SQL gives me better performance if I only 
>> need a
>> particular property of a node in the result set.
> can you elaborate on this? it would be great to have more background
> information on this...
> 
> afaik, there should not be any difference in performance, and for example
> "//element(*, nt:resource)/(@jcr:lastModified)" should be equivalent to
> "select jcr:lastModified from nt:resource"

Hmm ;) My fault. I made two assumptions that proved to be wrong ;p

1. I was blindly assuming that I have to use a NodeIterator to iterate over an 
XPath query result.
2. I assumed that RowIterator would only load the PropertyState of the 
properties in the SQL query and not the associated NodeStates.

This led me to using SQL queries with RowIterator. Your astonishment made me 
having a look at the implementation, where I found both assumptions to be wrong.

Just forget about the performance argument. And I will correct my answer to the 
survey, since I only need to use XPath now ;)

Thanks & Cheers,
Christoph


Mime
View raw message