Question

我从Lucene.Net 4.8演示项目（https://github.com/synhershko/LuceneNetDemo）开始。我的目标是能够使用查询解析器（QueryParser或MultiFieldQueryParser）来搜索文本和数字。那可能吗？我找到的只是使用范围（NumericRangeQuery）的示例，或建立我自己的查询解析器的建议。我无法确定是否可以通过现有的查询解析器创建范围？

using System;
using Lucene.Net.Store;
using Lucene.Net.Documents;
using Lucene.Net.Index;
using Lucene.Net.Util;
using Lucene.Net.QueryParsers.Classic;
using Lucene.Net.Search;
using Lucene.Net.Analysis.Standard;

/*
Package Manager:
Install-Package Lucene.Net -Version 4.8.0-beta00004 -Pre
Install-Package Lucene.Net.Analysis.Common -Version 4.8.0-beta00004 -Pre
Install-Package Lucene.Net.QueryParser -Version 4.8.0-beta00004 -Pre
*/

namespace LuceneNetNumbers
{
    class Program
    {
        static void Main(string[] args)
        {
            LuceneVersion MatchVersion = LuceneVersion.LUCENE_48;

            using (var oDirectory = new RAMDirectory())
            {
                var oAnalyzer = new StandardAnalyzer(MatchVersion);
                var oQueryParser = new MultiFieldQueryParser(MatchVersion, new[] { "name", "height", "age" }, oAnalyzer);
                var oIndexWriterConfig = new IndexWriterConfig(MatchVersion, oAnalyzer);
                var oIndexWriter = new IndexWriter(oDirectory, oIndexWriterConfig);
                var oSearcherManager = new SearcherManager(oIndexWriter, true, null);

                var oAdd = new Action<string, double, int>((sName, nAge, nHeight) =>
                {
                    var oDocument = new Document
                    {
                        new TextField("name", sName, Field.Store.YES),
                        new Int32Field("height", nHeight, Field.Store.YES),
                        new DoubleField("age", nAge, Field.Store.YES),
                    };

                    oIndexWriter.UpdateDocument(new Term("name", sName), oDocument);
                });

                oAdd("John Doe", 24.45, 56);
                oAdd("John Smith", 44.44, 64);
                oAdd("Mike Smith", 56.65, 70);

                oIndexWriter.Flush(true, true);
                oIndexWriter.Commit();

                //

                var oSearch = new Action<string>((sQueryString) =>
                {
                    var oQuery = oQueryParser.Parse(sQueryString);
                    oSearcherManager.MaybeRefreshBlocking();
                    var oSearcher = oSearcherManager.Acquire();

                    try
                    {
                        var oTopDocs = oSearcher.Search(oQuery, 10);
                        var nTotalHits = oTopDocs.TotalHits;
                        Console.WriteLine("Total Hits: {0}", nTotalHits);

                        foreach (var oResult in oTopDocs.ScoreDocs)
                        {
                            var oDocument = oSearcher.Doc(oResult.Doc);

                            var nScore = oResult.Score;
                            var sName = oDocument.GetField("name")?.GetStringValue();
                            var nAge = oDocument.GetField("age")?.GetNumericValue();
                            var nHeight = oDocument.GetField("height")?.GetNumericValue();

                            Console.WriteLine("{0:0.00}, {1,15}, {2,8}, {3,8}", nScore, sName, nAge, nHeight);
                        }
                    }
                    catch (Exception e)
                    {
                        Console.WriteLine(e.ToString());
                    }
                    finally
                    {
                        oSearcherManager.Release(oSearcher);
                        oSearcher = null;
                    }
                });

                oSearch("john");
                oSearch("height:64");

                /*
                Output:
                Total Hits: 2
                0.20,        John Doe,    24.45,       56
                0.20,      John Smith,    44.44,       64
                Total Hits: 0
                */
            }
        }
    }
}

Answer 1

它与Lucene.Net保存数值（编码形式）的方式有关：

new Int32Field("height", nHeight, Field.Store.YES)

您可以使用NumericRangeQuery

var oQuery = NumericRangeQuery.NewInt32Range("height", 64, 64, true, true);

使用您尝试搜索min和max值的号码。

另一种选择是使用TermQuery并将您的号码转换为BytesRef：

BytesRef bytes = new BytesRef(NumericUtils.BUF_SIZE_INT32);
NumericUtils.Int32ToPrefixCoded(64, 0, bytes);
Term term = new Term("height", bytes);
var oQuery = new TermQuery(term);

当然，您无法将查询解析为字符串，但您可以随时创建自己的解析器来组合术语：

BytesRef bytes = new BytesRef(NumericUtils.BUF_SIZE_INT32);
NumericUtils.Int32ToPrefixCoded(64, 0, bytes);
Term term = new Term("height", bytes);
// var oQuery = new TermQuery(term);

var oQuery = new BooleanQuery
{
    { new TermQuery(new Term("name", "John")), Occur.SHOULD },
    { new TermQuery(term), Occur.SHOULD }
};

您可以看到查询如何翻译为字符串

Answer 2

虽然我对我的想法并不完全满意，但它确实满足了我使用查询解析器搜索数字的要求......以及原始示例所暗示的文本。我将继续研究如何使用带有字符串和数字的StandardQueryParser.SetMultiFields和StandardQueryParser.NumericConfigMap，并在此处编辑/发布任何发现。

using Lucene.Net.Analysis.Standard;
using Lucene.Net.Documents;
using Lucene.Net.Index;
using Lucene.Net.QueryParsers.Flexible.Standard;
using Lucene.Net.QueryParsers.Flexible.Standard.Config;
using Lucene.Net.Search;
using Lucene.Net.Store;
using Lucene.Net.Support;
using Lucene.Net.Util;
using System;
using System.Globalization;

/*
Package Manager:
Install-Package Lucene.Net -Version 4.8.0-beta00004 -Pre
Install-Package Lucene.Net.Analysis.Common -Version 4.8.0-beta00004 -Pre
Install-Package Lucene.Net.QueryParser -Version 4.8.0-beta00004 -Pre
*/

namespace LuceneNetNumbers
{
    class Program
    {
        static void Main(string[] args)
        {
            LuceneVersion MatchVersion = LuceneVersion.LUCENE_48;

            using (var oDirectory = new RAMDirectory())
            {
                var oAnalyzer = new StandardAnalyzer(MatchVersion);

                //##########
                //##########

                //List of changes...

                //1. Remove this.
                //var oQueryParser = new MultiFieldQueryParser(MatchVersion, new[] { "name", "height", "age" }, oAnalyzer);

                //2. Add the following 6 lines of code.
                var oQueryParser = new StandardQueryParser(oAnalyzer);

                var oNumericConfigMap = new HashMap<string, NumericConfig>();
                oNumericConfigMap.Put("height", new NumericConfig(8, new NumberFormatIgnoreExceptions(CultureInfo.CurrentCulture), NumericType.INT32));
                oNumericConfigMap.Put("age", new NumericConfig(8, new NumberFormatIgnoreExceptions(CultureInfo.CurrentCulture), NumericType.DOUBLE));
                oQueryParser.NumericConfigMap = oNumericConfigMap;

                oQueryParser.SetMultiFields(new[] { "name", "height", "age" });

                //3. Add null as second parameter to StandardQueryParser.Parse below to utilize StandardQueryParser.SetMultiFields

                //4. Create NumberFormatIgnoreExceptions. I was not able to find another way (yet) to get 
                //StandardQueryParser.SetMultiFields and StandardQueryParser.NumericConfigMap to work with 
                //both text and number fields.  I feel like this is a bit of a hack, but it does satisfiy my
                //requirement of using a query parser to search for numbers (and text... implied by example).

                //##########
                //##########

                var oIndexWriterConfig = new IndexWriterConfig(MatchVersion, oAnalyzer);
                var oIndexWriter = new IndexWriter(oDirectory, oIndexWriterConfig);
                var oSearcherManager = new SearcherManager(oIndexWriter, true, null);

                var oAdd = new Action<string, double, int>((sName, nAge, nHeight) =>
                {
                    var oDocument = new Document
                    {
                        new TextField("name", sName, Field.Store.YES),
                        new Int32Field("height", nHeight, Field.Store.YES),
                        new DoubleField("age", nAge, Field.Store.YES),
                    };

                    oIndexWriter.UpdateDocument(new Term("name", sName), oDocument);
                });

                oAdd("John Doe", 24.45, 56);
                oAdd("John Smith", 44.44, 64);
                oAdd("Mike Smith", 56.65, 70);

                oIndexWriter.Flush(true, true);
                oIndexWriter.Commit();

                //

                var oSearch = new Action<string>((sQueryString) =>
                {
                    var oQuery = oQueryParser.Parse(sQueryString, null);
                    oSearcherManager.MaybeRefreshBlocking();
                    var oSearcher = oSearcherManager.Acquire();

                    try
                    {
                        var oTopDocs = oSearcher.Search(oQuery, 10);
                        var nTotalHits = oTopDocs.TotalHits;
                        Console.WriteLine("Total Hits: {0}", nTotalHits);

                        foreach (var oResult in oTopDocs.ScoreDocs)
                        {
                            var oDocument = oSearcher.Doc(oResult.Doc);

                            var nScore = oResult.Score;
                            var sName = oDocument.GetField("name")?.GetStringValue();
                            var nAge = oDocument.GetField("age")?.GetNumericValue();
                            var nHeight = oDocument.GetField("height")?.GetNumericValue();

                            Console.WriteLine("{0:0.00}, {1,15}, {2,8}, {3,8}", nScore, sName, nAge, nHeight);
                        }
                    }
                    catch (Exception e)
                    {
                        Console.WriteLine(e.ToString());
                    }
                    finally
                    {
                        oSearcherManager.Release(oSearcher);
                        oSearcher = null;
                    }
                });

                oSearch("john");
                oSearch("height:64");
                oSearch("age:[44.45 TO 56.66]");
                oSearch("height:[70 TO *]");

                /*
                Output:
                Total Hits: 2
                0.12,        John Doe,    24.45,       56
                0.12,      John Smith,    44.44,       64
                Total Hits: 1
                1.00,      John Smith,    44.44,       64
                Total Hits: 1
                1.00,      Mike Smith,    56.65,       70
                Total Hits: 1
                1.00,      Mike Smith,    56.65,       70
                */
            }
        }
    }

    class NumberFormatIgnoreExceptions : NumberFormat
    {
        public NumberFormatIgnoreExceptions(CultureInfo locale) : base(locale)
        {
        }

        public override object Parse(string source)
        {
            var oValue = default(object);

            try { oValue = base.Parse(source); } catch { }

            return oValue;
        }
    }
}

可以通过Lucene.Net（3.0.3或4.8.0）QueryParser搜索数字吗？

2 个答案: