跨度搜索SpanQuery【续】

上一篇文章讲了ApanTermQuery的用法，接下来继续：

SpanFirstQuery：

 

IndexSearcher searcher  =   new  IndexSearcher(path); 
                
                Term t1  =   new  Term( " content " ,  " david " ); 
                SpanTermQuery query1 = new  SpanTermQuery(t1);
                
                System.out.println( " 将Span的值设为3： " );
                SpanFirstQuery query = new  SpanFirstQuery(query1, 3 );
                
                Hits hits = searcher.search(query);
                 for ( int  i = 0 ;i < hits.length();i ++ )
                {
                    System.out.println( " ============ " );
                    System.out.println(hits.doc(i));
                    System.out.println( " ------------ " );
                }

注：构建SpanFirstQuery对象时，传入的参数不是一个词条Term，而是一个SpanTermQuery的对象实例。3表示从first开始到第3个位置内是否有要检索的内容。

SpanNearQuery：

 注意到定义“将Slop定义为2时，要重新写变量，必须吗？不需要要的。看看下面这个实例：

 

IndexSearcher searcher  =   new  IndexSearcher(path); 
                
                Term t1  =   new  Term( " content " ,  " david " ); 
                Term t2 = new  Term( " content " , " beautiful " );
                
                SpanTermQuery query1 = new  SpanTermQuery(t1);
                SpanTermQuery query2 = new  SpanTermQuery(t2);
                
                System.out.println( " 将Slop的值设为3： " );
                
                SpanNearQuery query = new  SpanNearQuery( new  SpanQuery[]{query1,query2}, 3 , false );
                
                Hits hits = searcher.search(query);
                 for ( int  i = 0 ;i < hits.length();i ++ )
                {
                    System.out.println( " ============ " );
                    System.out.println(hits.doc(i));
                    System.out.println( " ------------ " );
                }
                
                System.out.println( " 将Slop的值设为2： " );
                query = new  SpanNearQuery( new  SpanQuery[]{query1,query2}, 1 , false );
                
                hits = searcher.search(query);
                 for ( int  i = 0 ;i < hits.length();i ++ )
                {
                    System.out.println( " ============ " );
                    System.out.println(hits.doc(i));
                    System.out.println( " ------------ " );
                }  

 我原来是重新定义SpanNearQuery及后面的Hits但是太麻烦了，就用了下面的方法：

System.out.println( " 将Slop的值设为2： " );
                query = new  SpanNearQuery( new  SpanQuery[]{query1,query2}, 1 , false );
                
                hits = searcher.search(query);
                 for ( int  i = 0 ;i < hits.length();i ++ )
                {
                    System.out.println( " ============ " );
                    System.out.println(hits.doc(i));
                    System.out.println( " ------------ " );
                }  

是不是特简单啊，省得麻烦！

其实SpanNearQuery与PhraseQuey有异曲同工之妙，发现了吧！

复杂点的例子（混合SpanNearQuery）

 

IndexWriter writer  =   new  IndexWriter(path,  new  StandardAnalyzer(),  true ); 
                Document doc1  =   new  Document(); 

                Field f1  =   new  Field( " content " ,  " aa bb cc dd ee ff gg hh ii jj kk " , 
                Field.Store.YES, Field.Index.TOKENIZED); 
                doc1.add(f1); 
                writer.addDocument(doc1); 
                writer.close(); 
                
                IndexSearcher searcher  =   new  IndexSearcher(path); 
//                  创建四个词条 
                Term t1  =   new  Term( " content " ,  " aa " ); 
                Term t2  =   new  Term( " content " ,  " cc " ); 
                
                Term t3  =   new  Term( " content " ,  " gg " ); 
                Term t4  =   new  Term( " content " ,  " kk " ); 

//                  创建四个SpanTermQuery 
//                  SpanTermQuery的功能与TermQuery一样,只不过是把文档中数据源中的词语位置都标记下来了 
                SpanTermQuery s1  =   new  SpanTermQuery(t1); 
                SpanTermQuery s2  =   new  SpanTermQuery(t2); 
                SpanTermQuery s3  =   new  SpanTermQuery(t3); 
                SpanTermQuery s4  =   new  SpanTermQuery(t4); 

//                  先创建两个SpanNearQuery，参数为上面的四个SpanTermQuery 
//                  SpanNearQuery的功能与PhraseQuery的功能类似,只不过PharseQuery查询的是具有一定无关 
//                  单词长度的查询，而SpanNearQuery所匹配不一定是短语，它还有可能把SpanQuery的查询结果作为查询 
//                  对象。下面的构造函数：第二个参数相当于坡度slop，规定查询数据源的第一个必须出现查询的对象， 
//                  第三个参数false，表示要查询的多个词组是否要按照顺序出现在数据源中 
                SpanNearQuery query1  =   new  SpanNearQuery( new  SpanQuery[] { s1, s2 },  1 , 
                 false ); 
                SpanNearQuery query2  =   new  SpanNearQuery( new  SpanQuery[] { s3, s4 },  3 , 
                 false ); 

//                  在创建一个SpanNearQuery，参数为上面两个SpanNearQuery 
//                  意义就是在以上两个SpanNearQuery的搜索的结果中，进行搜索 
                SpanNearQuery query3  =   new  SpanNearQuery( new  SpanNearQuery[] { query1, query2 },  3 ,  false ); 

                Hits hits  =  searcher.search(query3); 
                 for  ( int  i  =   0 ; i   <  hits.length(); i ++ ) 
                { 
                     System.out.println(hits.doc(i)); 
                } 
                searcher.close(); 

形象图解如下：我不会弄啊！

                                                  Slop=3  

                    aa bb cc                dd ee ff                gg  hh  ii  j  kk

                    

                 SpanNearQuery                           SpanNearQuery

 

                                        SpanNearQuery

SpanOrQuery与SpanNotQuery那就太简单了在上面的例子中加入如下两行即可

SpanOrQuery是合并SpanQuery[]的搜索结果 
                    SpanOrQuery query4  =   new  SpanOrQuery( new  SpanQuery[] { query1, query2 }); 

//                      SpanNotQuery意思是从第一个SpanQuery的查询结果中去掉第二个SpanQuery的查询结果 
                    SpanNotQuery query5  =   new  SpanNotQuery(query1, query2);