atom feed26 messages in org.apache.lucene.java-devRe: FST and FieldCache?
FromSent OnAttachments
Davi...@MITRE.org)May 18, 2011 9:52 pm 
Dawid WeissMay 18, 2011 11:30 pm 
Earwin BurrfootMay 19, 2011 1:24 am 
Dawid WeissMay 19, 2011 1:42 am 
Michael McCandlessMay 19, 2011 3:08 am 
Dawid WeissMay 19, 2011 3:16 am 
Michael McCandlessMay 19, 2011 3:21 am 
Dawid WeissMay 19, 2011 3:36 am 
Earwin BurrfootMay 19, 2011 5:30 am 
Michael McCandlessMay 19, 2011 5:44 am 
Dawid WeissMay 19, 2011 5:45 am 
Earwin BurrfootMay 19, 2011 6:02 am 
Robert MuirMay 19, 2011 6:04 am 
Dawid WeissMay 19, 2011 6:20 am 
Jason RutherglenMay 19, 2011 6:22 am 
Earwin BurrfootMay 19, 2011 6:29 am 
Jason RutherglenMay 19, 2011 6:36 am 
Michael McCandlessMay 19, 2011 6:40 am 
Jason RutherglenMay 19, 2011 7:08 am 
Davi...@MITRE.org)May 19, 2011 7:53 am 
Davi...@MITRE.org)May 19, 2011 8:58 am 
Michael McCandlessMay 19, 2011 8:59 am 
Jason RutherglenMay 19, 2011 9:35 am 
Michael McCandlessMay 19, 2011 9:42 am 
Earwin BurrfootMay 19, 2011 11:00 am 
Dawid WeissMay 19, 2011 11:48 am 
Subject:Re: FST and FieldCache?
From:Dawid Weiss (dawi@cs.put.poznan.pl)
Date:May 19, 2011 6:20:37 am
List:org.apache.lucene.java-dev

I tried to prevent another reference to mr. Daciuk :)

Why? He's a very nice guy :)

But you have to lookup this err..index somehow. And that's either binary or hash lookup. Where's the win?

You can do a sparse O(1) index and have a slight gain from these few large initial states. This only makes sense if you perform tons of these lookups, really.

Mike's right -- the FST will output a structure that is ready to be used for by-number retrieval of strings (or anything else) as long as the numbers are strictly monotonous (and preferably continuous). The output will be what you're suggesting, Earwin -- accumulated sums.

Dawid