lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ahmet Arslan <iori...@yahoo.com.INVALID>
Subject Re: How to improve the performance in Lucene when query is long?
Date Tue, 11 Nov 2014 13:22:30 GMT
Hi Harry,

May be you can use BooleanQuery#setMinimumNumberShouldMatch method. What happens when you
use set it to half of the numTerms?

ahmet


On Tuesday, November 11, 2014 8:35 AM, Harry Yu <502437499@qq.com> wrote:
Hi everyone,



I have been using Lucene to build a POI searching & geocoding system. After test, I found
that when query is long(above 10 terms). And the speed of searching is too slow near to 1s.
I think the bottleneck is that I used OR to generate my BooleanQuery. It would get plenty
of candidates documents. And it would also consume too many time to score and rank.

I changed to use AND to generate my BooleanQuery. But it decrease the accuracy of hits. So
I want to find a solution to reduce candidate documents and do not decrease the accuracy in
this situation.

Thanks for your help!‍



------------------
Harry YuInstitute of Remote Sensing and Geographic Information System.
School of Earth and Space Sciences, Peking University;
Beijing, China, 100871;
Email: 502437499@qq.com OR harryyu1018@163.com 

---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org


Mime
View raw message