首先创建一个简单的java工程
创建lib目录并引入jar包
其中IKAnalyzer是一个第三方的中文分词器
要使用IK分词器还需要在根目录下引入几个文件
IKAnalyzer.cfg.xml
IK Analyzer 扩展配置
mydict.dic;
stopword.dic;
mydict.dic
不是坏人
奥巴马
普京
斯诺登
stopword.dic
a
an
and
are
as
创建索引存放目录
基本环境搭建完成
新建一个类 先写好几个公用的静态方法
//索引目录
public static String indexDir = "D:\\software\\idea-workspace\\lucene\\src\\main\\resources\\index_dir";
//获取分词器
private static Analyzer getAnalyzer() {
return new IKAnalyzer();
}
//获取FSDirectory
private static FSDirectory getFSDirectory() throws IOException {
FSDirectory directory = FSDirectory.open(new File(indexDir));
return directory;
}
//获取IndexWriter
private static IndexWriter getIndexWriter() throws IOException {
IndexWriterConfig conf = new IndexWriterConfig(Version.LATEST, getAnalyzer());
IndexWriter writer = new IndexWriter(getFSDirectory(), conf);
return writer;
}
接下来做简单增删查改操作
创建索引
public static void createIndex() throws IOException {
IndexWriter writer = getIndexWriter();
try {
Document doc = new Document();
doc.add(new StringField("id", "1", Field.Store.YES)); //不分词,索引,要存储
doc.add(new TextField("name", "张三", Field.Store.YES)); //分词,索引,存储
doc.add(new IntField("age", 23, Field.Store.YES)); //不分词,索引,存储
doc.add(new StoredField("icon", "http://icon/****")); //不分词,不索引,要存储
doc.add(new TextField("introduction", "个人介绍", Field.Store.NO));//分词,索引,不存储
writer.addDocument(doc);
} finally {
writer.close();
}
}
//main方法中调用
public static void main(String[] args) throws IOException {
createIndex();
}
索引目录中 产生了一些索引文件
term查询文档
//term查询文档
public static void query(String id) throws IOException {
//new TermQuery(new Term(SITE_ID, siteId.toString()))
IndexReader reader = DirectoryReader.open(getFSDirectory());
try {
IndexSearcher searcher = new IndexSearcher(reader);
Query query = new TermQuery(new Term("id", id));
//n代表查出前10条
TopDocs topDocs = searcher.search(query, 10);
System.out.println(">>>总记录数" + topDocs.totalHits);
ScoreDoc[] docs = topDocs.scoreDocs;
for (ScoreDoc scoreDoc : docs) {
int docId = scoreDoc.doc; //文档的id 跟存储的id没什么关系
Document doc = searcher.doc(docId);//获取doc对象
System.out.println(doc.get("id"));
System.out.println(doc.get("name"));
System.out.println(doc.get("age"));
System.out.println(doc.get("icon"));
System.out.println(doc.get("introduction"));
}
} finally {
reader.close();
}
}
执行main函数
public static void main(String[] args) throws IOException {
// createIndex();
query("1");
}
输出如下
可以看到introduction的输出为null 这就是前面创建Field的时候 指定不存储
>>>总记录数1
1
张三
23
http://icon/****
null
term删除
其实是根据查询条件删除
public static void delete(String id) throws IOException {
IndexWriter indexWriter = getIndexWriter();
try {
indexWriter.deleteDocuments(new Term("id", id));
// indexWriter.deleteAll(); 删除所有文档
} finally {
indexWriter.close(); //close的时候才会真正的删除
}
}
main
public static void main(String[] args) throws IOException {
// createIndex();
delete("1");
query("1");
}
更新索引
public static void updateIndex(String id,String name) throws IOException {
IndexWriter writer = getIndexWriter();
try {
Document doc = new Document();
doc.add(new StringField("id", id, Field.Store.YES)); //不分词,索引,要存储
doc.add(new TextField("name", name, Field.Store.YES)); //分词,索引,存储
writer.updateDocument(new Term("id", id), doc);
} finally {
writer.close();
}
}
main函数
public static void main(String[] args) throws IOException {
createIndex();
// delete("1");
query("1");
updateIndex("1","李四");
query("1");
}
可以看到其实也是先删除了 然后再添加
>>>总记录数1
1
张三
23
http://icon/****
null
>>>总记录数1
1
李四
null
null
null
BooleanQuery查询文档 并指定排序 懵B的 先记录下来吧
public static void booleanQuery(String queryString) throws IOException {
//new TermQuery(new Term(SITE_ID, siteId.toString()))
IndexReader reader = DirectoryReader.open(getFSDirectory());
IndexSearcher searcher = new IndexSearcher(reader);
try {
BooleanQuery bq = new BooleanQuery();
Query q;
q = MultiFieldQueryParser.parse(queryString, new String[]{"name", "introduction"},
new BooleanClause.Occur[]{BooleanClause.Occur.SHOULD,
BooleanClause.Occur.SHOULD}, getAnalyzer());
bq.add(q, BooleanClause.Occur.MUST); //且
//查询年龄 age>=1 && age <=30
q = NumericRangeQuery.newIntRange("age", 1, 30, true, true);
bq.add(q, BooleanClause.Occur.MUST); //且
Sort sort = new Sort();
List sortFieldList = new ArrayList();
//按照id从小到大
sortFieldList.add(new SortField("id", SortField.Type.STRING, false));
//并且 按照年龄从大到小
sortFieldList.add(new SortField("age", SortField.Type.INT, true));
sort.setSort(sortFieldList.toArray(new SortField[sortFieldList.size()]));
//n代表查出前10条
TopDocs topDocs = searcher.search(bq, 10, sort);
System.out.println(">>>总记录数" + topDocs.totalHits);
ScoreDoc[] docs = topDocs.scoreDocs;
for (ScoreDoc scoreDoc : docs) {
int docId = scoreDoc.doc; //文档的id 跟存储的id没什么关系
Document doc = searcher.doc(docId);//获取doc对象
System.out.println(doc.get("id"));
System.out.println(doc.get("name"));
System.out.println(doc.get("age"));
System.out.println(doc.get("icon"));
System.out.println(doc.get("introduction"));
}
} catch (ParseException e) {
e.printStackTrace();
} finally {
reader.close();
}
}
自定义评分也记录下
/**
* 重写评分的实现方式
**/
private class MyScoreProvider extends CustomScoreProvider {
private AtomicReaderContext context;
public MyScoreProvider(AtomicReaderContext context) {
super(context);
this.context = context;
}
/**
* 重写评分方法
**/
@Override
public float customScore(int doc, float subQueryScore, float valSrcScore) throws IOException {
// 从域缓存中加载索引字段信息
FieldCache.Ints ageInts = FieldCache.DEFAULT.getInts(context.reader(), "age", false);
int age = ageInts.get(doc);
//这里可以通过 一些 文档Field的值 来为文档打分 达到排序的效果
//doc实际上就是Lucene中得docId
float score = subQueryScore * valSrcScore; //分数
// 判断加权
//....
/*
* 此处可以控制与原有得分结合的方式，加减乘除都可以
* **/
return score;
}
}
/**
* 重写CustomScoreQuery 的getCustomScoreProvider方法
* 引用自定义的Provider
*/
private class MyCustomScoreQuery extends CustomScoreQuery {
public MyCustomScoreQuery(Query subQuery) {
super(subQuery);
}
@Override
protected CustomScoreProvider getCustomScoreProvider(AtomicReaderContext context) throws IOException {
/**注册使用自定义的评分实现方式**/
return new MyScoreProvider(context);
}
}
查询语句换一下 这里不能指定sort参数 sort参数优先级高
TopDocs docs = searcher.search(new MyCustomScoreQuery(query),10);