平台:Lucene 2.1.0,JRE 1.4,Oracle 10g,IBM Web Sphere。
数据表:Article。字段:ID(自动增长),Title(String),Content(String)。共有550000条记录。
对Article建立索引:
1import org.apache.lucene.analysis.*;
2import org.apache.lucene.analysis.cn.*;
3import org.apache.lucene.document.*;
4import org.apache.lucene.index.*;
5import java.sql.*;
6import oracle.jdbc.pool.*;
7
8public class Index {
9 private String url="jdbc:oracle:thin:@//192.168.0.l:1521/Test";
10 private String user="terry";
11 private String password="dev";
12 private Connection con=null;
13 private Statement st=null;
14 private ResultSet rs=null;
15 private String indexUrl="E:\\ArticleIndex";
16
17 private ResultSet getResult() throws Exception{
18 OracleDataSource ods=new OracleDataSource();
19
20 ods.setURL(this.url);
21 ods.setUser(this.user);
22 ods.setPassword(this.password);
23
24 this.con=ods.getConnection();
25 this.st=this.con.createStatement();
26 this.rs=this.st.executeQuery("SELECT * FROM Article");
27
28 return this.rs;
29 }
30
31 public void createIndex() throws Exception{
32 ResultSet rs=this.getResult();
33
34 Analyzer chineseAnalyzer=new ChineseAnalyzer();
35 IndexWriter indexWriter=new IndexWriter(this.indexUrl,chineseAnalyzer,true);
36 indexWriter.setMergeFactor(100);
37 indexWriter.setMaxBufferedDocs(100);
38
39 java.util.Date startDate=new java.util.Date();
40
41 System.out.println("开始索引时间:"+startDate);
42
43 executeIndex(rs,indexWriter);
44
45 indexWriter.optimize();
46
47 indexWriter.close();
48
49 java.util.Date endDate=new java.util.Date();
50
51 System.out.println("索引结束时间:"+endDate);
52 System.out.println("共花费:"+(endDate.getTime()-startDate.getTime())+"ms");
53 }
54
55 private void executeIndex(ResultSet rs,IndexWriter indexWriter) throws Exception{
56 int i=0;
57
58 while(rs.next()){
59 int id=rs.getInt("ID");
60 String title=rs.getString("TITLE");
61 String info=rs.getString("CONTENT");
62
63 Document doc=new Document();
64
65 Field idField=new Field("ID",Integer.toString(id),Field.Store.YES,Field.Index.NO,Field.TermVector.NO);
66 Field titleField=new Field("Title",title,Field.Store.YES,Field.Index.TOKENIZED,Field.TermVector.YES);
67 Field infoField=new Field("Content",title,Field.Store.YES,Field.Index.TOKENIZED,Field.TermVector.YES);
68
69 doc.add(idField);
70 doc.add(titleField);
71 doc.add(infoField);
72
73 indexWriter.addDocument(doc);
74
75 i++;
76 }
77
78 this.close();
79
80 System.out.println("共处理记录:"+i);
81 }
82
83 private void close() throws Exception{
84 this.rs.close();
85 this.st.close();
86 this.con.close();
87 }
88}
查找:
1import java.io.*;
2import org.apache.lucene.analysis.cn.*;
3import org.apache.lucene.search.*;
4import org.apache.lucene.store.*;
5import org.apache.lucene.document.*;
6import org.apache.lucene.queryParser.QueryParser;
7
8import java.util.*;
9
10public class Search {
11
12 private static final String indexUrl="E:\\ArticleIndex";
13
14 public static void main(String[] args) throws Exception {
15/**//*建立索引代码,查找时注释*/
16 //Index index=new Index();
17
18 //index.createIndex();
19
20
21
22
23 File indexDir=new File(indexUrl);
24 FSDirectory fdir=FSDirectory.getDirectory(indexDir);
25
26 IndexSearcher searcher=new IndexSearcher(fdir);
27
28//对中文建立解析(必须)
29 QueryParser parser=new QueryParser("Title",new ChineseAnalyzer());
30 Query query=parser.parse("李湘");
31
32 Date startDate=new Date();
33 System.out.println("检索开始时间:"+startDate);
34
35 Hits result=searcher.search(query);
36
37 for(int i=0;i<result.length();i++){
38 Document doc=result.doc(i);
39
40 System.out.println("内容:"+doc.get("Content"));
41 }
42
43 Date endDate=new Date();
44
45 System.out.println("共有记录:"+result.length());
46 System.out.println("共花费:"+(endDate.getTime()-startDate.getTime()));
47 }
48
49}
经测试,建立索引文件大概花了11分钟。一般情况下,和用SQL执行LIKE查询差不多。
当然,这只是我的粗略测试。最近一阶段,我会对Lucene进行代码深入研究。
posted on 2007-04-30 16:43
Terry Liang 阅读(2135)
评论(3) 编辑 收藏 所属分类:
Lucene 2.1研究