Lee -
 
I'm pretty sure that unordered isn't going to help you much.  It's
implementation-dependent, and our ordering tends to be optimized as you
outlined. 
 
ian

________________________________

From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Lee
Pollington
Sent: Monday, April 30, 2007 8:14 AM
To: General Mark Logic Developer Discussion
Subject: [MarkLogic Dev General] unordered


Hi,

A general question, I'm not trying to solve a problem. 

What does ML do with unordered( <expr> ) ? I've run a couple of simple
tests - a straight pull of a 1000 nodes, 1200 nodes based on IDs in
another document (with warm caches)  and I see little or no difference. 

My understanding is that the fragment IDs from the fact lists will
already be in database order. Any range indexes will require a sort to
simplify the intersection with any additional fragment ID lists, so
there is little scope for unordered to have an impact on performance.

Are there any scenarios where unordered is *likely* to provide a
performance benefit? 

Thanks,
Lee

_______________________________________________
General mailing list
[email protected]
http://xqzone.com/mailman/listinfo/general

Reply via email to