lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Uwe Schindler (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (LUCENE-3096) MultiSearcher does not work correctly with Not on NumericRange
Date Sat, 14 May 2011 07:21:47 GMT

    [ https://issues.apache.org/jira/browse/LUCENE-3096?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13033478#comment-13033478
] 

Uwe Schindler commented on LUCENE-3096:
---------------------------------------

This is a well-known bug (LUCENE-2756), which is unfixable (query rewrite across different
searchers is wrong) without totally changing the way how queries are rewritten.

To fix the bug, you should use a MultiReader on your IndexReaders and use a simple IndexSearcher
on top of that MultiReader:

{code}
IndexReader[] readers;
readers[0] = IndexReader.open(directory);
readers[1] = IndexReader.open(otherdirectory);
...
IndexSearcher searcher = new IndexSearcher(new MultiReader(readers));
{code}

MultiSearcher and ParallelMultiSearcher were deprecated in 3.1 because of this and disappear
in coming Lucene 4.0. ParallelMultiSearcher functionality is now available through IndexSearcher
in 3.1 (it parallelizes across index segments, LUCENE-2837).

I will close this as won't fix if nobody objects.

> MultiSearcher does not work correctly with Not on NumericRange
> --------------------------------------------------------------
>
>                 Key: LUCENE-3096
>                 URL: https://issues.apache.org/jira/browse/LUCENE-3096
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Search
>    Affects Versions: 3.0.2
>            Reporter: John Wang
>
> Hi, Keith
> My colleague xiaoyang and I just confirmed that this is actually due to a lucene bug
on Multisearcher. In particular,
> If we search with Not on NumericRange and we use MultiSearcher, we
> will wrong search results (However, if we use IndexSearcher, the
> result is correct).  Basically the NotOfNumericRange does not have
> impact on multisearcher. We suspect it is because the createWeight()
> function in MultiSearcher and hope you can help us to fix this bug of
> lucene. I attached the code to reproduce this case. Please check it
> out.
> In the attached code, I have two separate functions :
> (1) testNumericRangeSingleSearcher(Query query)
>     where I create 6 documents, with a field called "id"= 1,2,3,4,5,6
> respectively . Then I search by the query which is
>     +MatchAllDocs -NumericRange(3,3). The expected result then should
> be 5 hits since the document 3 is MUST_NOT.
> (2) testNumericRangeMultiSearcher(Query query)
>     where i create 2 RamDirectory(), each of which has 3 documents,
> 1,2,3; and 4,5,6. Then I search by the same query as above using
> multiSearcher. The expected result should also be 5 hits.
> However, from (1), we get 5 hits = expected results, while in (2) we
> get 6 hits != expected results.
> We also experimented this with our zoie/bobo open source tools and get
> the same results because our multi-bobo-browser is built on
> multi-searcher in lucene.
> I already emailed the lucene community group. Hopefully we can get some feedback soon.
> If you have any further concern, pls let me know! 
> Thank you very much!
> Code:  (based on lucene 3.0.x)
> import java.io.IOException;
> import java.io.PrintStream;
> import java.text.DecimalFormat;
> import org.apache.lucene.analysis.WhitespaceAnalyzer;
> import org.apache.lucene.document.Document;
> import org.apache.lucene.document.Field;
> import org.apache.lucene.document.NumericField;
> import org.apache.lucene.index.CorruptIndexException;
> import org.apache.lucene.index.IndexWriter;
> import org.apache.lucene.index.Term;
> import org.apache.lucene.search.BooleanQuery;
> import org.apache.lucene.search.FieldCache;
> import org.apache.lucene.search.IndexSearcher;
> import org.apache.lucene.search.MatchAllDocsQuery;
> import org.apache.lucene.search.MultiSearcher;
> import org.apache.lucene.search.NumericRangeQuery;
> import org.apache.lucene.search.Query;
> import org.apache.lucene.search.ScoreDoc;
> import org.apache.lucene.search.Searchable;
> import org.apache.lucene.search.Sort;
> import org.apache.lucene.search.SortField;
> import org.apache.lucene.search.TermQuery;
> import org.apache.lucene.search.TopDocs;
> import org.apache.lucene.search.BooleanClause.Occur;
> import org.apache.lucene.store.Directory;
> import org.apache.lucene.store.LockObtainFailedException;
> import org.apache.lucene.store.RAMDirectory;
> import com.convertlucene.ConvertFrom2To3;
> public class TestNumericRange
> {
>  public final static void main(String[] args)
>  {
>    try
>    {
>      BooleanQuery query = new  BooleanQuery();
>      query.add(NumericRangeQuery.newIntRange("numId", 3, 3, true,
> true), Occur.MUST_NOT);
>      query.add(new MatchAllDocsQuery(), Occur.MUST);
>      testNumericRangeSingleSearcher(query);
>      testNumericRangeMultiSearcher(query);
>    }
>    catch(Exception e)
>    {
>      e.printStackTrace();
>    }
>  }
>  public static void testNumericRangeSingleSearcher(Query query)
> throws CorruptIndexException, LockObtainFailedException, IOException
>  {
>     String[] ids = {"1", "2", "3", "4", "5", "6"};
>    Directory directory = new RAMDirectory();
>    IndexWriter writer = new IndexWriter(directory, new
> WhitespaceAnalyzer(),  IndexWriter.MaxFieldLength.UNLIMITED);
>    for (int i = 0; i < ids.length; i++)
>    {
>      Document doc = new Document();
>      doc.add(new Field("id", ids[i],
>                        Field.Store.YES,
>                        Field.Index.NOT_ANALYZED));
>      doc.add(new NumericField("numId").setIntValue(Integer.valueOf(ids[i])));
>      writer.addDocument(doc);
>    }
>    writer.close();
>    IndexSearcher searcher = new IndexSearcher(directory);
>    TopDocs docs = searcher.search(query, 10);
>    System.out.println("SingleSearcher: testNumericRange: hitNum: " +
> docs.totalHits);
>    for(ScoreDoc doc : docs.scoreDocs)
>    {
>      System.out.println(searcher.explain(query, doc.doc));
>    }
>    searcher.close();
>    directory.close();
>  }
>  public static void testNumericRangeMultiSearcher(Query query) throws
> CorruptIndexException, LockObtainFailedException, IOException
>  {
>     String[] ids1 = {"1", "2", "3"};
>    Directory directory1 = new RAMDirectory();
>    IndexWriter writer1 = new IndexWriter(directory1, new
> WhitespaceAnalyzer(),  IndexWriter.MaxFieldLength.UNLIMITED);
>    for (int i = 0; i < ids1.length; i++)
>    {
>      Document doc = new Document();
>      doc.add(new Field("id", ids1[i],
>                        Field.Store.YES,
>                        Field.Index.NOT_ANALYZED));
>      doc.add(new NumericField("numId").setIntValue(Integer.valueOf(ids1[i])));
>      writer1.addDocument(doc);
>    }
>    writer1.close();
>    String[] ids2 = {"4", "5", "6"};
>    Directory directory2 = new RAMDirectory();
>    IndexWriter writer2 = new IndexWriter(directory2, new
> WhitespaceAnalyzer(),  IndexWriter.MaxFieldLength.UNLIMITED);
>    for (int i = 0; i < ids2.length; i++)
>    {
>      Document doc = new Document();
>      doc.add(new Field("id", ids2[i],
>                        Field.Store.YES,
>                        Field.Index.NOT_ANALYZED));
>      doc.add(new NumericField("numId").setIntValue(Integer.valueOf(ids2[i])));
>      writer2.addDocument(doc);
>    }
>    writer2.close();
>    IndexSearcher[] searchers = new IndexSearcher[2];
>    searchers[0] = new IndexSearcher(directory1);
>    searchers[1] = new IndexSearcher(directory2);
>    MultiSearcher multiSearcher = new MultiSearcher(searchers);
>    TopDocs docs = multiSearcher.search(query, 10);
>    System.out.println("MultiSearcher: testNumericRange: hitNum: " +
> docs.totalHits);
>    for(ScoreDoc doc : docs.scoreDocs)
>    {
>      System.out.println(multiSearcher.explain(query, doc.doc));
>    }
>    multiSearcher.close();
>    directory1.close();
>    directory2.close();
>  }
> }

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org


Mime
View raw message