Lucene的使用
如果你想快速查询你磁盘上文件,或查询邮件、Web页面,甚至查询存于数据库的数据,你都可以借助于Lucene来完成。但是要完成查询就必须先建立索引。首先从Lucene API说起:
1、 Lucene API(核心操作类)
IndexWriter | 创建和维护索引(向原索引中添加新Document,设置合并策略、优化等) |
FSDirectory | 最主要用来存储索引文件的类,表示将索引文件存储到文件系统 |
Document | 索引和查询的原子单元,一个Document包含一系列Field |
IndexReader | 一个抽象类,提供了访问索引的接口,当然访问索引也可以通过它的子类来完成 |
Analyzer | 分词类,它有一系列子类,都是用来将文本解析成TokenStream |
Searcher | 用于查询索引的核心类 |
2、创建索引
Directory dir = FSDirectory.open(new File("lucene.blog")); IndexWriter writer = new IndexWriter(dir,new StandardAnalyzer(Version.LUCENE_29),true, IndexWriter.MaxFieldLength.UNLIMITED); Document doc = new Document(); doc.add(new Field("id", "101", Field.Store.YES, Field.Index.NO)); doc.add(new Field("name", "kobe bryant", Field.Store.YES, Field.Index.NO)); writer.addDocument(doc); writer.optimize(); writer.close();
Directory dir = FSDirectory.open(new File("lucene.blog")); IndexWriter writer = new IndexWriter(dir,new StandardAnalyzer(Version.LUCENE_29),true, IndexWriter.MaxFieldLength.UNLIMITED); Document doc = new Document(); doc.add(new Field("id", "101", Field.Store.YES, Field.Index.NO)); doc.add(new Field("name", "kobe bryant", Field.Store.YES, Field.Index.NO)); writer.addDocument(doc); writer.optimize(); writer.close();
如上所示将索引文件存储于工作目录下lucene.blog文件夹 ,创建了Document,向Document里添加了两个Field id和name,然后使用IndexWriter的addDocument(Document)方法将其添加到索引目录下的索引文件中,然后使用IndexWriter的optimize()方法进行对索引文件优化,最后关闭IndexWriter;
3、通过IndexWriter删除索引中Document
Directory dir = FSDirectory.open(new File("lucene.blog")); IndexWriter writer = new IndexWriter(dir, new StandardAnalyzer(Version.LUCENE_29), true, IndexWriter.MaxFieldLength.UNLIMITED); writer.deleteDocuments(new Term("id", "101")); writer.commit(); writer.close();
Directory dir = FSDirectory.open(new File("lucene.blog")); IndexWriter writer = new IndexWriter(dir, new StandardAnalyzer(Version.LUCENE_29), true, IndexWriter.MaxFieldLength.UNLIMITED); writer.deleteDocuments(new Term("id", "101")); writer.commit(); writer.close();
如上先打开索引位置(工作目录下lucene.blog文件夹 ),然后直接调运IndexWriter的deleteDocuments(Term)方法删除上面2中创建的Document,注意必须调运commit()方法,上面2中之所以没有commit()是因为optimize()方法中存在默认Commit方法;
4、通过IndexWriter更新索引中Document
Directory dir = FSDirectory.open(new File("lucene.blog")); IndexWriter writer = new IndexWriter(dir, new StandardAnalyzer(Version.LUCENE_29), true, IndexWriter.MaxFieldLength.UNLIMITED); Document doc = new Document(); doc.add(new Field("id", "101", Field.Store.YES, Field.Index.ANALYZED)); // Field.Index.ANALYZED doc.add(new Field("name", "kylin soong", Field.Store.YES, Field.Index.ANALYZED)); writer.updateDocument(new Term("id", "101"), doc); writer.commit(); writer.close();
Directory dir = FSDirectory.open(new File("lucene.blog")); IndexWriter writer = new IndexWriter(dir, new StandardAnalyzer(Version.LUCENE_29), true, IndexWriter.MaxFieldLength.UNLIMITED); Document doc = new Document(); doc.add(new Field("id", "101", Field.Store.YES, Field.Index.NO)); doc.add(new Field("name", "kylin soong", Field.Store.YES, Field.Index.NO)); writer.updateDocument(new Term("id", "101"), doc); writer.commit(); writer.close();
通过IndexWriter的updateDocument(Term, Document)来完成更新,具体是将包含Term("id", "101")的Document删除,然后将传入的Document添加到索引文件;
5、Field选项意义
Field field = new Field( "101", "kobe bryant", Field.Store.YES, Field.Index.ANALYZED, Field.TermVector.YES);
Field field = new Field( "101", "kobe bryant", Field.Store.YES, Field.Index.ANALYZED, Field.TermVector.YES);
如上代码显示Field各属性设置情况,下面简单说明这些属性选项的意义
Field.Store.*决定是否将Field的完全值进行存储,注意:不能将整个文本内容存储,这样导致索引文件过大
Field.Store.YES | 存储,一旦存储,你可以用完整的Field的完全值作为查询条件查询(id:101) |
Field.Store.NO | 不存储 |
Field.Index.*控制Field的值是否可查询通过索引成的索引文件
Field.Index.ANALYZED | 用Analyzer将Field的值分词成多个Token |
Field.Index.NOT_ANALYZED | 不对Field的值分词,将Field的值作为一个Token处理 |
Field.Index.ANALYZED_NO_NORMS | 类似ANALYZED,但不存常规信息到索引文件 |
Field.Index.NOT_ANALYZED_NO_NORMS | 类似NOT_ANALYZED,但不存常规信息到索引文件 |
Field.Index.NO | 不进行索引,Field的值不可被搜索 |
如果你想要检索出唯一的terms在搜索时,或对搜索结果进行加亮处理等操作是Field.TermVector.*是必要的
Field.TermVector.YES | 记录唯一的terms,当重复发生时记下重复数,在不做额外处理 |
Field.TermVector.WITH_POSITIONS | 在上面基础上记录下位置 |
Field.TermVector.WITH_OFFSETS | 在TermVector.YES基础上记录偏移量 |
Field.TermVector.WITH_POSITIONS_OFFSETS | 在TermVector.YES基础上记录偏移量和位置 |
Field.TermVector.NO | 不做任何处理 |
6、索引numbers
Document doc = new Document(); NumericField field1 = new NumericField("id"); field1.setIntValue(101); doc.add(field1); NumericField field2 = new NumericField("price"); field1.setDoubleValue(123.50); doc.add(field2);
Document doc = new Document(); NumericField field1 = new NumericField("id"); field1.setIntValue(101); doc.add(field1); NumericField field2 = new NumericField("price"); field1.setDoubleValue(123.50); doc.add(field2);
如上所示为索引numbers方法;
7、索引Date和Time
Document doc = new Document(); doc.add(new NumericField("timestamp").setLongValue(new Date().getTime())); doc.add(new NumericField("day").setIntValue((int) (new Date().getTime()/24/3600))); Calendar cal = Calendar.getInstance(); cal.setTime(new Date()); doc.add(new NumericField("dayOfMonth").setIntValue(cal.get(Calendar.DAY_OF_MONTH)));
Document doc = new Document(); doc.add(new NumericField("timestamp").setLongValue(new Date().getTime())); doc.add(new NumericField("day").setIntValue((int) (new Date().getTime()/24/3600))); Calendar cal = Calendar.getInstance(); cal.setTime(new Date()); doc.add(new NumericField("dayOfMonth").setIntValue(cal.get(Calendar.DAY_OF_MONTH)));
实质上对Date和Time的处理是将Date和Time转化为numbers来处理,注意:当然也可以把Date和Time以及上面的numbers当做字符串来处理,不过这样影响查询;
8、IndexWriter的其他同法
Directory dir = FSDirectory.open(new File("lucene.blog")); IndexWriter writer = new IndexWriter(dir, new StandardAnalyzer(Version.LUCENE_29), true, IndexWriter.MaxFieldLength.LIMITED); writer.setMaxFieldLength(1); MergePolicy policy = new LogByteSizeMergePolicy(writer); writer.setMergePolicy(policy); writer.optimize(5); writer.close();
Directory dir = FSDirectory.open(new File("lucene.blog")); IndexWriter writer = new IndexWriter(dir, new StandardAnalyzer(Version.LUCENE_29), true, IndexWriter.MaxFieldLength.LIMITED); writer.setMaxFieldLength(1); MergePolicy policy = new LogByteSizeMergePolicy(writer); writer.setMergePolicy(policy); writer.optimize(5); writer.close();
如上IndexWriter.MaxFieldLength.LIMITED设定了Field截取功能,如果Field值相当长,而你只想索引Field值的前固定个字符,可以用Field截取功能来实现;IndexWriter的setMergePolicy(policy),可以设定合并策略,另外optimize(int maxNumSegments)方法可以通过参数设定优化成的Segment个数;
9、根据确定的term查询
IndexReader reader = IndexReader.open(FSDirectory.open(new File("lucene.blog")),true); IndexSearcher searcher = new IndexSearcher(reader); Term term = new Term("id","101"); Query query = new TermQuery(term); TopDocs topDocs = searcher.search(query, 10); System.out.println(topDocs.totalHits); ScoreDoc[] docs = topDocs.scoreDocs; System.out.println(docs[0].doc + " " + docs[0].score); Document doc = searcher.doc(docs[0].doc); System.out.println(doc.get("id"));
IndexReader reader = IndexReader.open(FSDirectory.open(new File("lucene.blog")),true); IndexSearcher searcher = new IndexSearcher(reader); Term term = new Term("id","101"); Query query = new TermQuery(term); TopDocs topDocs = searcher.search(query, 10); System.out.println(topDocs.totalHits); ScoreDoc[] docs = topDocs.scoreDocs; System.out.println(docs[0].doc + " " + docs[0].score); Document doc = searcher.doc(docs[0].doc); System.out.println(doc.get("id"));
如上示例显示了一个Lucene查询的基本方法,IndexSearcher是核心的查询类,IndexReader 可以读取索引文件,IndexSearcher有一系列重载的Search()方法,可以根据传入不同参数进行不同查询处理,ScoreDoc数组保存查询结果,和相关得分;
10、根据QueryParser查询,并收集查询结果
IndexReader reader = IndexReader.open(FSDirectory.open(new File("lucene.blog")),true); IndexSearcher searcher = new IndexSearcher(reader); Analyzer analyzer = new StandardAnalyzer(Version.LUCENE_29); QueryParser parser = new QueryParser(Version.LUCENE_29,"name",analyzer); String queryString = "kobe"; Query query = parser.parse(queryString); TopScoreDocCollector collector = TopScoreDocCollector.create(10, false); searcher.search(query, collector); ScoreDoc[] hits = collector.topDocs().scoreDocs; for(int i = 0 ; i < hits.length ; i ++) { Document doc = searcher.doc(hits[i].doc); String name = doc.get("name"); if (name != null) { System.out.println(name); } }
IndexReader reader = IndexReader.open(FSDirectory.open(new File("lucene.blog")),true); IndexSearcher searcher = new IndexSearcher(reader); Analyzer analyzer = new StandardAnalyzer(Version.LUCENE_29); QueryParser parser = new QueryParser(Version.LUCENE_29,"name",analyzer); String queryString = "kobe"; Query query = parser.parse(queryString); TopScoreDocCollector collector = TopScoreDocCollector.create(10, false); searcher.search(query, collector); ScoreDoc[] hits = collector.topDocs().scoreDocs; for(int i = 0 ; i < hits.length ; i ++) { Document doc = searcher.doc(hits[i].doc); String name = doc.get("name"); if (name != null) { System.out.println(name); } }
如上为一个使用QueryParser查询关键字“kobe”的实例,另外还对查询结果进行了收集
11、使用Lucene图形化工具Luke来操作索引
Luke使用非常简单:
下载:http://code.google.com/p/luke/ 点击下载最新版本,下载完成直接点击下载的jar包,就可以进入图形化操作界面,选择索引的目录就可以对索引进行图形化操作
相关推荐
renjinlong 2020-09-03
Jacry 2020-07-04
IceStreamLab 2020-06-26
mengyue 2020-06-09
PasserbyX 2020-05-16
mameng 2020-05-12
心丨悦 2020-05-06
编码之路 2020-05-03
mengyue 2020-05-02
qiuzhuoxian 2020-02-23
编码之路 2020-02-20
lionelf 2020-02-03
TyCoding 2020-02-01
heniancheng 2020-01-31
某某某 2020-01-30
PinkBean 2020-01-29
某某某 2020-01-12
编码之路 2020-01-01
itmale 2020-01-01