I debugged your test and the column-label map is never used because the query 
only has two columns.

Regards
Carl

From: Noel Grandin [mailto:[email protected]]
Sent: den 26 april 2012 13:34
To: [email protected]
Cc: Carl Hasselskog
Subject: Re: Potential performance improvement

Hmm, I ran the measurements myself, and I'm seeing an improvement, but it's on 
the order of 0.1%

What might be more interesting is to see how much of a difference it makes to 
your own application, since your tests indicated that it was showing up on the 
performance profile.

On 2012-04-26 13:22, Noel Grandin wrote:
Cool, thanks for your work.

Can you take measurements, then apply this patch to the test code, and report 
on the difference?

On 2012-04-26 13:05, Carl Hasselskog wrote:
I've now created a patch for this (see attachment). Feel free to use it under 
whatever license suits you.

When running then benchmark the difference was negligible. However, after 
reading the benchmark code I couldn't find any part of that actually would 
benefit from this patch (since all read are by column-index not by name). 
Perhaps that it something that could be added (e.g. by making queryReadResult 
read by column-name)? My guess would be that getting by column-name would be 
more common so that could potentially make the tests more similar to a 
real-world scenario.

Regards
Carl


-- 
You received this message because you are subscribed to the Google Groups "H2 
Database" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/h2-database?hl=en.

Reply via email to