lucene 查询的使用

  1 package cn.itcast.i_query;
  2 
  3 import java.util.ArrayList;
  4 import java.util.List;
  5 
  6 import org.apache.lucene.document.Document;
  7 import org.apache.lucene.index.Term;
  8 import org.apache.lucene.search.BooleanQuery;
  9 import org.apache.lucene.search.FuzzyQuery;
 10 import org.apache.lucene.search.IndexSearcher;
 11 import org.apache.lucene.search.MatchAllDocsQuery;
 12 import org.apache.lucene.search.NumericRangeQuery;
 13 import org.apache.lucene.search.Query;
 14 import org.apache.lucene.search.TermQuery;
 15 import org.apache.lucene.search.TopDocs;
 16 import org.apache.lucene.search.WildcardQuery;
 17 import org.apache.lucene.search.BooleanClause.Occur;
 18 import org.junit.Test;
 19 
 20 import cn.itcast._domain.Article;
 21 import cn.itcast._util.ArticleDocumentUtils;
 22 import cn.itcast._util.LuceneUtils;
 23 
 24 public class TestApp {
 25 
 26     // 关键词查询
 27     @Test
 28     public void testTermQuery() {
 29         // 对应的查询字符串为：title:lucene
 30         TermQuery query = new TermQuery(new Term("title", "lucene"));
 31         searchAndShowResult(query);
 32     }
 33 
 34     // 通配符查询
 35     // ? 表示一个任意字符
 36     // * 表示0或多个任意字符
 37     @Test
 38     public void testWildcardQuery() {
 39         // 对应的查询字符串为：title:lu*n?
 40         // WildcardQuery query = new WildcardQuery(new Term("title", "lu*n?"));
 41 
 42         // 对应的查询字符串为：content:互?网
 43         WildcardQuery query = new WildcardQuery(new Term("content", "互?网"));
 44         searchAndShowResult(query);
 45     }
 46 
 47     // 查询所有
 48     @Test
 49     public void testMatchAllDocsQuery() {
 50         // 对应的查询字符串为：*:*
 51         MatchAllDocsQuery query = new MatchAllDocsQuery();
 52         searchAndShowResult(query);
 53     }
 54 
 55     // 模糊查询
 56     @Test
 57     public void testFuzzyQuery() {
 58         // 对应的查询字符串为：title:lucenX~0.9
 59         // 第二个参数是最小相似度，表示有多少正确的就显示出来，比如0.9表示有90%正确的字符就会显示出来。
 60         FuzzyQuery query = new FuzzyQuery(new Term("title", "lucenX"), 0.8F);
 61         searchAndShowResult(query);
 62     }
 63 
 64     // 范围查询
 65     @Test
 66     public void testNumericRangeQuery() {
 67         // 对应的查询字符串为：id:[5 TO 15]
 68         // NumericRangeQuery query = NumericRangeQuery.newIntRange("id", 5, 15, true, true);
 69 
 70         // 对应的查询字符串为：id:{5 TO 15}
 71         // NumericRangeQuery query = NumericRangeQuery.newIntRange("id", 5, 15, false, false);
 72 
 73         // 对应的查询字符串为：id:[5 TO 15}
 74         NumericRangeQuery query = NumericRangeQuery.newIntRange("id", 5, 15, true, false);
 75 
 76         searchAndShowResult(query);
 77     }
 78 
 79     // 布尔查询
 80     @Test
 81     public void testBooleanQuery() {
 82         BooleanQuery booleanQuery = new BooleanQuery();
 83         // booleanQuery.add(query, Occur.MUST); // 必须满足
 84         // booleanQuery.add(query, Occur.SHOULD); // 多个SHOULD一起用表示OR的关系
 85         // booleanQuery.add(query, Occur.MUST_NOT); // 非
 86 
 87         Query query1 = new TermQuery(new Term("title", "lucene"));
 88         Query query2 = NumericRangeQuery.newIntRange("id", 5, 15, false, true);
 89 
 90         // // 对应的查询字符串为：+title:lucene +id:{5 TO 15]
 91         // // 对应的查询字符串为（大写的AND）：title:lucene AND id:{5 TO 15]
 92         // booleanQuery.add(query1, Occur.MUST);
 93         // booleanQuery.add(query2, Occur.MUST);
 94 
 95         // 对应的查询字符串为：title:lucene id:{5 TO 15]
 96         // 对应的查询字符串为：title:lucene OR id:{5 TO 15]
 97         // booleanQuery.add(query1, Occur.SHOULD);
 98         // booleanQuery.add(query2, Occur.SHOULD);
 99 
100         // 对应的查询字符串为：+title:lucene -id:{5 TO 15]
101         // 对应的查询字符串为：title:lucene (NOT id:{5 TO 15] )
102         booleanQuery.add(query1, Occur.MUST);
103         booleanQuery.add(query2, Occur.MUST_NOT);
104 
105         searchAndShowResult(booleanQuery);
106     }
107 
108     /**
109      * 测试搜索的工具方法
110      * 
111      * @param query
112      */
113     private void searchAndShowResult(Query query) {
114         try {
115             // // 准备查询条件
116             // String queryString = "content:lucene";
117             // // 1，把查询字符串转为Query对象（从title和content中查询）
118             // QueryParser queryParser = new MultiFieldQueryParser(Version.LUCENE_30, new String[] { "title", "content" }, LuceneUtils.getAnalyzer());
119             // Query query = queryParser.parse(queryString);
120 
121             System.out.println("--->  // 对应的查询字符串为：" + query + "\n");
122 
123             // 2，执行查询，得到中间结果
124             IndexSearcher indexSearcher = new IndexSearcher(LuceneUtils.getDirectory()); // 指定所用的索引库
125             TopDocs topDocs = indexSearcher.search(query, 100); // 最多返回前n条结果
126 
127             // 3，处理结果
128             List<Article> list = new ArrayList<Article>();
129             for (int i = 0; i < topDocs.scoreDocs.length; i++) {
130                 // 根据编号拿到Document数据
131                 int docId = topDocs.scoreDocs[i].doc; // Document的内部编号
132                 Document doc = indexSearcher.doc(docId);
133                 // 把Document转为Article
134                 Article article = ArticleDocumentUtils.documentToArticle(doc);
135                 list.add(article);
136             }
137             indexSearcher.close();
138 
139             // 显示结果
140             System.out.println("总结果数：" + list.size());
141             for (Article a : list) {
142                 System.out.println("------------------------------");
143                 System.out.println("id = " + a.getId());
144                 System.out.println("title = " + a.getTitle());
145                 System.out.println("content = " + a.getContent());
146             }
147         } catch (Exception e) {
148             throw new RuntimeException(e);
149         }
150     }
151 }