|
|
Forget.... ;-) I use 32-bit platform (Fedora Core release 3), 1 GB RAM,
Celeron 2.40GHz.
-----Original Message-----
From: Erik Hatcher [mailto:erik@xxxxxxxxxxxxxxxxxxxxx]
Sent: Friday, March 10, 2006 4:33 PM
To: java-user@xxxxxxxxxxxxxxxxx
Subject: Re: speed
Importance: High
I'm curious, how many documents in your index and what kind of
queries are you issuing? Even 1.6 seconds is a LONG time for a
search in most Lucene indexes.
Erik
On Mar 10, 2006, at 5:38 AM, Anton Potehin wrote:
> I try 2 way for get count results:
>
> 1)
> Query q = ....
> IndexSearcher is = ....
> int count = Is.search(q).length();
>
> 2)
> Query q = ....
> IndexSearcher is = ....
> CountCollector collector = new CountCollector();
> Is.search(q, collector);
> Int count = collector.getCount();
>
> First way return results for 1.644 sec.
> Second way return results fot 3.088 sec.
> Why first way faster then second?
> Can I accelerate it? And how? Can you offer any ways for get number of
> results?
>
>
> ----------------------CountCollector
> import org.apache.lucene.search.HitCollector;
>
> /**
> * @author Anton Potekhin
> * @date: 03.03.2006 17:28:03
> */
> public class CountCollector extends HitCollector {
>
> private int numberRusults = 0;
> public int getCount() {
> return numberRusults;
> }
>
> public CountCollector() {
>
> }
>
> public void collect(int doc, float score) {
> if (score > 0.0f) {
> numberRusults++;
> }
> }
> }
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: java-user-unsubscribe@xxxxxxxxxxxxxxxxx
> For additional commands, e-mail: java-user-help@xxxxxxxxxxxxxxxxx
---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@xxxxxxxxxxxxxxxxx
For additional commands, e-mail: java-user-help@xxxxxxxxxxxxxxxxx
---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@xxxxxxxxxxxxxxxxx
For additional commands, e-mail: java-user-help@xxxxxxxxxxxxxxxxx
|
|