Lucene.net系列六 -- search 下

本文主要结合测试案例介绍了Lucene下的各种查询语句以及它们的简化方法.

通过本文你将了解Lucene的基本查询语句,并通过学习相关的测试代码以加强了解.

源代码下载

具体的查询语句

在了解了SQL后, 你是否想了解一下查询语法树?在这里简要介绍一些能被Lucene直接使用的查询语句.

1.         TermQuery
查询某个特定的词,在文章开始的例子中已有介绍.常用于查询关键字.

              [Test]
         public void Keyword()
         {
              IndexSearcher searcher = new IndexSearcher(directory);
              Term t = new Term("isbn", "1930110995");
              Query query = new TermQuery(t);
              Hits hits = searcher.Search(query);
              Assert.AreEqual(1, hits.Length(), "JUnit in Action");
         }

注意Lucene中的关键字,是需要用户去保证唯一性的.

TermQuery和QueryParse

只要在QueryParse的Parse方法中只有一个word,就会自动转换成TermQuery.

2.         RangeQuery
用于查询范围,通常用于时间,还是来看例子:

namespace dotLucene.inAction.BasicSearch
{
     public class RangeQueryTest : LiaTestCase
     {
         private Term begin, end;

         [SetUp]
         protected override void Init()
         {
              begin = new Term("pubmonth", "200004");

              end = new Term("pubmonth", "200206");
              base.Init();
         }

         [Test]
         public void Inclusive()
         {
              RangeQuery query = new RangeQuery(begin, end, true);
              IndexSearcher searcher = new IndexSearcher(directory);

              Hits hits = searcher.Search(query);
              Assert.AreEqual(1, hits.Length());
         }

         [Test]
         public void Exclusive()
         {
              RangeQuery query = new RangeQuery(begin, end, false);
              IndexSearcher searcher = new IndexSearcher(directory);

              Hits hits = searcher.Search(query);
              Assert.AreEqual(0, hits.Length());
         }

     }
}

RangeQuery的第三个参数用于表示是否包含该起止日期.

RangeQuery 和 QueryParse

               [Test]
         public void TestQueryParser()
         {
              Query query = QueryParser.Parse("pubmonth:[200004 TO 200206]", "subject", new SimpleAnalyzer());
              Assert.IsTrue(query is RangeQuery);
              IndexSearcher searcher = new IndexSearcher(directory);
              Hits hits = searcher.Search(query);

              query = QueryParser.Parse("{200004 TO 200206}", "pubmonth", new SimpleAnalyzer());
              hits = searcher.Search(query);
              Assert.AreEqual(0, hits.Length(), "JDwA in 200206");
         }

Lucene用[] 和{}分别表示包含和不包含.

3.    PrefixQuery

用于搜索是否包含某个特定前缀,常用于Catalog的检索.

             [Test]
         public void TestPrefixQuery()
         {
              PrefixQuery query = new PrefixQuery(new Term("category", "/Computers"));

             IndexSearcher searcher = new IndexSearcher(directory);
              Hits hits = searcher.Search(query);
              Assert.AreEqual(2, hits.Length());
             
              query = new PrefixQuery(new Term("category", "/Computers/JUnit"));
              hits = searcher.Search(query);
              Assert.AreEqual(1, hits.Length(), "JUnit in Action");
         }

PrefixQuery和QueryParse

            [Test]
         public void TestQueryParser()
         {

              QueryParser qp = new QueryParser("category", new SimpleAnalyzer());
              qp.SetLowercaseWildcardTerms(false);
              Query query =qp.Parse("/Computers*");
              Console.Out.WriteLine("query = {0}", query.ToString());
              IndexSearcher searcher = new IndexSearcher(directory);
              Hits hits = searcher.Search(query);
              Assert.AreEqual(2, hits.Length());
              query =qp.Parse("/Computers/JUnit*");
              hits = searcher.Search(query);
              Assert.AreEqual(1, hits.Length(), "JUnit in Action");
         }

这里需要注意的是我们使用了QueryParser对象,而不是QueryParser类. 原因在于使用对象可以对QueryParser的一些默认属性进行修改.比如在上面的例子中我们的category是大写的,而QueryParser默认会把所有的含*的查询字符串变成小写/computer*. 这样我们就会查不到原文中的/Computers* ,所以我们需要通过设置QueryParser的默认属性来改变这一默认选项.即qp.SetLowercaseWildcardTerms(false)所做的工作.

4.     BooleanQuery

用于测试满足多个条件.

下面两个例子用于分别测试了满足与条件和或条件的情况.

         [Test]
         public void And()
         {
              TermQuery searchingBooks =
                   new TermQuery(new Term("subject", "junit"));

              RangeQuery currentBooks =
                   new RangeQuery(new Term("pubmonth", "200301"),
                                  new Term("pubmonth", "200312"),
                                  true);
              BooleanQuery currentSearchingBooks = new BooleanQuery();
              currentSearchingBooks.Add(searchingBooks, true, false);
              currentSearchingBooks.Add(currentBooks, true, false);
              IndexSearcher searcher = new IndexSearcher(directory);
              Hits hits = searcher.Search(currentSearchingBooks);

              AssertHitsIncludeTitle(hits, "JUnit in Action");
         }
         [Test]
         public void Or()
         {
              TermQuery methodologyBooks = new TermQuery(
                   new Term("category",
                            "/Computers/JUnit"));
              TermQuery easternPhilosophyBooks = new TermQuery(
                   new Term("category",
                            "/Computers/Ant"));
              BooleanQuery enlightenmentBooks = new BooleanQuery();
              enlightenmentBooks.Add(methodologyBooks, false, false);
              enlightenmentBooks.Add(easternPhilosophyBooks, false, false);
              IndexSearcher searcher = new IndexSearcher(directory);
              Hits hits = searcher.Search(enlightenmentBooks);
              Console.Out.WriteLine("or = " + enlightenmentBooks);
              AssertHitsIncludeTitle(hits, "Java Development with Ant");
              AssertHitsIncludeTitle(hits, "JUnit in Action");

         }

什么时候是与什么时候又是或? 关键在于BooleanQuery对象的Add方法的参数.

参数一是待添加的查询条件.

参数二Required表示这个条件必须满足吗? True表示必须满足, False表示可以不满足该条件.

参数三Prohibited表示这个条件必须拒绝吗? True表示这么满足这个条件的结果要排除, False表示可以满足该条件.

这样会有三种组合情况,如下表所示:

BooleanQuery 和 QueryParse

         [Test]
         public void TestQueryParser()
         {
              Query query = QueryParser.Parse("pubmonth:[200301 TO 200312] AND junit", "subject", new SimpleAnalyzer());
              IndexSearcher searcher = new IndexSearcher(directory);
              Hits hits = searcher.Search(query);
              Assert.AreEqual(1, hits.Length());
              query = QueryParser.Parse("/Computers/JUnit OR /Computers/Ant", "category", new WhitespaceAnalyzer());
              hits = searcher.Search(query);
              Assert.AreEqual(2, hits.Length());
         }

注意AND和OR的大小 如果想要Ａ与非B 就用 A AND �CB 表示, +A �CB也可以.

默认的情况下QueryParser会把空格认为是或关系,就象google一样.但是你可以通过QueryParser对象修改这一属性.

[Test]
         public void TestQueryParserDefaultAND()
         {
              QueryParser qp = new QueryParser("subject", new SimpleAnalyzer());
               qp.SetOperator(QueryParser.DEFAULT_OPERATOR_AND );
              Query query = qp.Parse("pubmonth:[200301 TO 200312] junit");
              IndexSearcher searcher = new IndexSearcher(directory);
              Hits hits = searcher.Search(query);
              Assert.AreEqual(1, hits.Length());

         }
5.         PhraseQuery
查询短语,这里面主要有一个slop的概念, 也就是各个词之间的位移偏差, 这个值会影响到结果的评分.如果slop为0,当然最匹配.看看下面的例子就比较容易明白了,有关slop的计算用户就不需要理解了,不过slop太大的时候对查询效率是有影响的,所以在实际使用中要把该值设小一点. PhraseQuery对于短语的顺序是不管的,这点在查询时除了提高命中率外,也会对性能产生很大的影响, 利用SpanNearQuery可以对短语的顺序进行控制,提高性能.
      [SetUp]
     protected void Init()
     {
         // set up sample document
         RAMDirectory directory = new RAMDirectory();
         IndexWriter writer = new IndexWriter(directory,
                                              new WhitespaceAnalyzer(), true);
         Document doc = new Document();
         doc.Add(Field.Text("field",
                            "the quick brown fox jumped over the lazy dog"));
         writer.AddDocument(doc);
         writer.Close();

         searcher = new IndexSearcher(directory);
     }
        private bool matched(String[] phrase, int slop)
     {
         PhraseQuery query = new PhraseQuery();
         query.SetSlop(slop);

         for (int i = 0; i < phrase.Length; i++)
         {
              query.Add(new Term("field", phrase[i]));
         }

         Hits hits = searcher.Search(query);
         return hits.Length() > 0;
     }

     [Test]
     public void SlopComparison()
     {
         String[] phrase = new String[]{"quick", "fox"};

         Assert.IsFalse(matched(phrase, 0), "exact phrase not found");

         Assert.IsTrue(matched(phrase, 1), "close enough");
     }

      [Test]
     public void Reverse()
     {
         String[] phrase = new String[] {"fox", "quick"};

         Assert.IsFalse(matched(phrase, 2), "exact phrase not found");

         Assert.IsTrue(matched(phrase, 3), "close enough");
     }

     [Test]
     public void Multiple()-
     {
         Assert.IsFalse(matched(new String[] {"quick", "jumped", "lazy"}, 3), "not close enough");
         Assert.IsTrue(matched(new String[] {"quick", "jumped", "lazy"}, 4), "just enough");
         Assert.IsFalse(matched(new String[] {"lazy", "jumped", "quick"}, 7), "almost but not quite");
         Assert.IsTrue(matched(new String[] {"lazy", "jumped", "quick"}, 8), "bingo");
     }

PhraseQuery和QueryParse

利用QueryParse进行短语查询的时候要先设定slop的值,有两种方式如下所示

[Test]
     public void TestQueryParser()
     {
         Query q1 = QueryParser.Parse(""quick fox"",
              "field", new SimpleAnalyzer());
         Hits hits1 = searcher.Search(q1);
         Assert.AreEqual(hits1.Length(), 0);

         Query q2 = QueryParser.Parse(""quick fox"~1",          //第一种方式
                                     "field", new SimpleAnalyzer());
         Hits hits2 = searcher.Search(q2);
         Assert.AreEqual(hits2.Length(), 1);

         QueryParser qp = new QueryParser("field", new SimpleAnalyzer());
         qp.SetPhraseSlop(1);                                    //第二种方式
         Query q3=qp.Parse(""quick fox"");
         Assert.AreEqual(""quick fox"~1", q3.ToString("field"),"sloppy, implicitly");
         Hits hits3 = searcher.Search(q2);
         Assert.AreEqual(hits3.Length(), 1);
     }

6.         WildcardQuery
通配符搜索,需要注意的是child, mildew的分值是一样的.
         [Test]
         public void Wildcard()
         {
              IndexSingleFieldDocs(new Field[]
                   {
                       Field.Text("contents", "wild"),
                       Field.Text("contents", "child"),
                       Field.Text("contents", "mild"),
                       Field.Text("contents", "mildew")
                   });
              IndexSearcher searcher = new IndexSearcher(directory);
              Query query = new WildcardQuery(
                   new Term("contents", "?ild*"));
              Hits hits = searcher.Search(query);
              Assert.AreEqual(3, hits.Length(), "child no match");
              Assert.AreEqual(hits.Score(0), hits.Score(1), 0.0, "score the same");
              Assert.AreEqual(hits.Score(1), hits.Score(2), 0.0, "score the same");
         }
WildcardQuery和QueryParse
需要注意的是出于性能的考虑使用QueryParse的时候,不允许在开头就使用就使用通配符.
同样处于性能考虑会将只在末尾含有*的查询词转换为PrefixQuery.
         [Test, ExpectedException(typeof (ParseException))]
         public void TestQueryParserException()
         {
              Query query = QueryParser.Parse("?ild*", "contents", new WhitespaceAnalyzer());
         }

         [Test]
         public void TestQueryParserTailAsterrisk()
         {
              Query query = QueryParser.Parse("mild*", "contents", new WhitespaceAnalyzer());
              Assert.IsTrue(query is PrefixQuery);
              Assert.IsFalse(query is WildcardQuery);

         }

         [Test]
         public void TestQueryParser()
         {
              Query query = QueryParser.Parse("mi?d*", "contents", new WhitespaceAnalyzer());
              Hits hits = searcher.Search(query);
              Assert.AreEqual(2, hits.Length());
         }
7.         FuzzyQuery
模糊查询, 需要注意的是两个匹配项的分值是不同的,这点和WildcardQuery是不同的

         [Test]
         public void Fuzzy()
         {
              Query query = new FuzzyQuery(new Term("contents", "wuzza"));
              Hits hits = searcher.Search(query);
              Assert.AreEqual( 2, hits.Length(),"both close enough");
              Assert.IsTrue(hits.Score(0) != hits.Score(1),"wuzzy closer than fuzzy");
              Assert.AreEqual("wuzzy", hits.Doc(0).Get("contents"),"wuzza bear");
         }

FuzzyQuery和QueryParse

注意和PhraseQuery中表示slop的区别,前者~后要跟数字.

         [Test]
         public void TestQueryParser()
         {
              Query query =QueryParser.Parse("wuzza~","contents",new SimpleAnalyzer());
              Hits hits = searcher.Search(query);
              Assert.AreEqual( 2, hits.Length(),"both close enough");
         }