【转载保存】lucene3.0可以对docId、docField、queryParser设置Boost值来影响排序结果

转自：http://catastiger.iteye.com/blog/803796

前提:不对结果做sort操作.
在搜索中,并不是所有的Document和Fields都是平等的.有些技术会要求到对其Doucment或者Fields的权值改变,默认值为:1.0F,以上需求都是通过改变Document的boost因子来改变的. 下面是通过lucene3.0,IKAnalyzer
1.通过设置doc boost改变排序结果

Java代码

/**
* 设置DOC boost 值影响查询排序结果
* @throws Exception
*/
public void testBoost1() throws Exception{
System.out.println("设置DOC boost 值影响查询排序结果");
RAMDirectory ramDir = new RAMDirectory();
Analyzer analyzer = new IKAnalyzer();
IndexWriter iw = new IndexWriter(ramDir, analyzer, true ,IndexWriter.MaxFieldLength.LIMITED);
String[] nameList = { "you are my friend", "a are my wife", "I love you" };
String[] addList = { "b", "you are my wife", "c" };
String[] fileList = { "1", "2", "3" };
for (int i = 0; i < nameList.length; i++){
Document doc = new Document();
doc.add(new Field("name", nameList[i], Field.Store.YES, Field.Index.ANALYZED));
doc.add(new Field("file", fileList[i], Field.Store.YES, Field.Index.ANALYZED));
doc.add(new Field("address", addList[i], Field.Store.YES, Field.Index.ANALYZED));
if (i == 2) {
doc.setBoost(2.0f);
}
// 这里设置了第三个文档优先级最高，所以在搜索出来的结果中，该文档排在最前
iw.addDocument(doc);
}
iw.close();
IndexSearcher _searcher = new IndexSearcher(ramDir);
String[] fields =new String[]{"name","address"};
Query query=IKQueryParser.parseMultiField(fields, "you");
TopDocs topDocs = _searcher.search(query,_searcher.maxDoc());
ScoreDoc[] hits = topDocs.scoreDocs;
for (int i = 0; i < hits.length; i++) {
Document doc = _searcher.doc(hits[i].doc);
System.out.println("name:"+doc.get("name"));
System.out.println("file:"+doc.get("file"));
}
_searcher.close();
}

if (i == 2) { doc.setBoost(2.0f); }这样I love you 将先输出，
2.通过设置query 影响排序

Java代码

/**
* 设置query boost值影响排序结果,如果有排序sort，则完全按照sort结果进行
* @throws Exception
*/
public void testBoost2() throws Exception{
System.out.println("设置query boost值影响排序结果");
RAMDirectory ramDir = new RAMDirectory();
Analyzer analyzer = new IKAnalyzer();
IndexWriter iw = new IndexWriter(ramDir, analyzer, true ,IndexWriter.MaxFieldLength.LIMITED);
String[] nameList = { "you are my friend", "a are my wife", "I love you" };
String[] addList = { "b", "you are my wife", "c" };
String[] fileList = { "1", "2", "3" };
for (int i = 0; i < nameList.length; i++)
{
Document doc = new Document();
doc.add(new Field("name", nameList[i], Field.Store.YES, Field.Index.ANALYZED));
doc.add(new Field("file", fileList[i], Field.Store.YES, Field.Index.ANALYZED));
doc.add(new Field("address", addList[i], Field.Store.YES, Field.Index.ANALYZED));
iw.addDocument(doc);
}
iw.close();
IndexSearcher _searcher = new IndexSearcher(ramDir);
BooleanQuery bq = new BooleanQuery();
QueryParser _parser = new QueryParser(Version.LUCENE_30,"name",analyzer);
Query _query = _parser.parse("you");
_query.setBoost(2f);
QueryParser _parser1 = new QueryParser(Version.LUCENE_30,"address",analyzer);
Query _query1 = _parser1.parse("you");
_query1.setBoost(1f);
bq.add(_query, BooleanClause.Occur.SHOULD);
bq.add(_query1, BooleanClause.Occur.SHOULD);
//
// for(int i=0;i<2;i++){
// QueryParser parser = new MultiFieldQueryParser(Version.LUCENE_30,new String[] {"name", "address" }, analyzer);
// Query q1 = parser.parse("you");
// bq.add(q1, BooleanClause.Occur.MUST);
// }
//
// SortField[] sortFields = new SortField[1];
// SortField sortField = new SortField("file", SortField.INT, true);//false升序，true降序
// sortFields[0] = sortField;
// Sort sort = new Sort(sortFields);
// TopDocs topDocs = _searcher.search(bq,null,_searcher.maxDoc(),sort);
//
TopDocs topDocs = _searcher.search(bq,_searcher.maxDoc());
ScoreDoc[] hits = topDocs.scoreDocs;
for (int i = 0; i < hits.length; i++) {
Document doc = _searcher.doc(hits[i].doc);
System.out.println("name:"+doc.get("name"));
System.out.println("file:"+doc.get("file"));
}
_searcher.close();
}

结果如下：（name 的boost最高，所以name优先于address排序在前面）
设置query boost值影响排序结果
name:you are my friend
file:1
name:I love you
file:3
name:a are my wife
file:2

3.通过设置fields 的boost 影响排序

Java代码

/**
* 设置field boost 值影响查询排序结果,有排序则按照排序
* @throws Exception
*/
//没设置field boost 213 设置后是132
public void testBoost3() throws Exception{
System.out.println("设置fields boost 值影响查询排序结果");
RAMDirectory ramDir = new RAMDirectory();
Analyzer analyzer = new IKAnalyzer();
IndexWriter iw = new IndexWriter(ramDir, analyzer, true ,IndexWriter.MaxFieldLength.LIMITED);
String[] nameList = { "you are my friend", "a are my wife", "I love you" };
String[] addList = { "b", "you are my wife", "c" };
String[] fileList = { "1", "2", "3" };
for (int i = 0; i < nameList.length; i++)
{
Document doc = new Document();
Field nameField = new Field("name", nameList[i], Field.Store.YES, Field.Index.ANALYZED);
nameField.setBoost(20f);
doc.add(nameField);
doc.add(new Field("file", fileList[i], Field.Store.YES, Field.Index.ANALYZED));
Field f = new Field("address", addList[i], Field.Store.YES, Field.Index.ANALYZED);
f.setBoost(30f);
doc.add(f);
iw.addDocument(doc);
}
iw.close();
IndexSearcher _searcher = new IndexSearcher(ramDir);
String[] fields =new String[]{"name","file","address"};
Query query=IKQueryParser.parseMultiField(fields, "you");
// SortField[] sortFields = new SortField[1];
// SortField sortField = new SortField("file", SortField.INT, true);//false升序，true降序
// sortFields[0] = sortField;
// Sort sort = new Sort(sortFields);
// TopDocs topDocs = _searcher.search(query,null,_searcher.maxDoc(),sort);
TopDocs topDocs = _searcher.search(query,_searcher.maxDoc());
ScoreDoc[] hits = topDocs.scoreDocs;
for (int i = 0; i < hits.length; i++) {
Document doc = _searcher.doc(hits[i].doc);
System.out.println("name:"+doc.get("name"));
System.out.println("file:"+doc.get("file"));
}
_searcher.close();
}

结果如下：（address 的boost最高，先排在前面了）
设置fields boost 值影响查询排序结果
name:a are my wife
file:2
name:you are my friend
file:1
name:I love you
file:3

【转载保存】lucene3.0可以对docId、docField、queryParser设置Boost值来影响排序结果相关推荐

Lucene-3.0.0配置
Lucene-3.0.0配置 2009-12-02 21:38 一.Lucene开发环境配置 step1.Lucene开发包下载 step2.Java开发环境配置 step3.Tomcat安装 ste ...
lucene3.0范围查找TermRangeQuery
原文链接:http://www.wenhq.com/article/view_415.html 欢迎转载,请注明出处:亲亲宝宝 lucene3.0范围查找TermRangeQuery 在lucene3 ...
lucene3.0中使用MultiFieldQueryParser多字段查找
使用Lucene做全文检索,一般我们经常会在多个字段(域)中查找,而不一定关心在那个字段中包含需要查找的值.比如在搜索框中输入:亲亲宝宝软件开发,只要标题.内容.作者等包含"亲亲宝宝软件 ...
[转载]IIS 6.0配置HTTP压缩的步骤
[转载]IIS 6.0配置HTTP压缩的步骤原文: http://zlei12.go.nease.net/ 1. HTTP压缩概述 HTTP压缩是在Web服务器和浏览器间传输压缩文本内容的方法.HT ...
[转载]AxureRP 7.0部件详解（一）
转载]AxureRP 7.0部件详解(一) 本文为Axure RT7.0教程,本章主要介绍menu菜单.table表格.Tree Widget 树部件三个部件,后续将持续更新...... Menu 菜 ...
mongodb 3.0.x 添加用户名和密码设置权限
mongodb 3.0.x 添加用户名,密码,权限设置最近遇到关于mongodb 3.0.x权限设置的问题,做了一下总结: 第一给mongodb 3.0.x 添加user 第一步修改配置文件: xx ...
InnoDB 的辅助索引叶子节点为什么不直接保存的记录地址而要存主键键值
转载自 InnoDB 的辅助索引叶子节点为什么不直接保存的记录地址而要存主键键值 RT,最近看书关于innodb的 InnoDB是索引组织表,所以完整的数据记录都存在聚集索引的叶子节点上辅助索引的 ...
ODOO v10.0 自动生成财务凭证的科目设置
ODOO v10.0 自动生成财务凭证的科目设置可以在产品类别及产品档案里设置,建议在产品类别下设置,方便维护. 项目设置为(具体科目以公司科目表为主) 对应作用业务 Income Account ...
显示一张桌子的信息，包括桌子的形状(长方形、方形、圆形、椭圆形；使用Rect、Square、Circle、Ellipse)、腿数、高度、桌面面积。定义变量来保存桌子的信息，并显示各个信息的值。要点提示
[问题描述] 显示一张桌子的信息,包括桌子的形状(长方形.方形.圆形.椭圆形:使用Rect.Square.Circle.Ellipse).腿数.高度.桌面面积.定义变量来保存桌子的信息,并显示各个信息 ...

【转载保存】lucene3.0可以对docId、docField、queryParser设置Boost值来影响排序结果

【转载保存】lucene3.0可以对docId、docField、queryParser设置Boost值来影响排序结果相关推荐

最新文章

热门文章