ZhongHua Wu created LUCENE-8767:
-----------------------------------

             Summary: DisjunctionMaxQuery do not work well when multiple search 
term+synonyms+mm+query fields with different fieldType.
                 Key: LUCENE-8767
                 URL: https://issues.apache.org/jira/browse/LUCENE-8767
             Project: Lucene - Core
          Issue Type: Bug
          Components: core/queryparser
    Affects Versions: 7.3
         Environment: Solr: 7.3.1

Backup:

FieldType for name field:

<fieldType name="forName" class="solr.TextField" positionIncrementGap="100" 
omitNorms="true">
 <analyzer>
 <tokenizer class="solr.WhitespaceTokenizerFactory"/>
 <filter class="solr.StopFilterFactory" ignoreCase="true" words="stopwords.txt" 
enablePositionIncrements="true" />
 <filter class="solr.WordDelimiterFilterFactory" generateWordParts="1" 
generateNumberParts="0" catenateWords="1" catenateNumbers="1" catenateAll="0" 
 splitOnCaseChange="0" preserveOriginal="1" splitOnNumerics="0"/>
 <filter class="solr.LowerCaseFilterFactory"/>
 <filter class="solr.SnowballPorterFilterFactory" language="English" 
protected="protwords.txt" />
 <filter class="solr.RemoveDuplicatesTokenFilterFactory"/>
 </analyzer>
 </fieldType>

FieldType for partNumber field:

<fieldType name="forPartNumber" class="solr.TextField" sortMissingLast="true" 
omitNorms="true">
 <analyzer>
 <tokenizer class="solr.KeywordTokenizerFactory"/>
 <filter class="solr.LowerCaseFilterFactory"/>
 <filter class="solr.TrimFilterFactory" />
 </analyzer>
 </fieldType>
            Reporter: ZhongHua Wu


When multiple fields in query fields came from different fieldType, especially 
one from KeywordTokenizerFactory, another from WhitespaceTokenizerFactory, then 
the generated parse query could not honor synonyms and mm, which hit incorrect 
documents. The following is my detail:
 # We use Solr 7.3.1
 # Our qf=name^10 partNumber_ntk, while fieldType of name use 
solr.WhitespaceTokenizerFactory and solr.WordDelimiterFilterFactory, while  
partNumber_ntk is not tokenized and use solr.KeywordTokenizerFactory
 # mm=2<3 4<5 6<-80%25
 # The search term is versatil sundress, while 'versatile' and 'testing' are 
synonyms, we have documents named " Versatil Empire Waist Sundress" which 
should be hit, but failed.
 # We test same query on Solr 5.5.4, it works fine, it do not work on Solr 
7.3.1.

q=

(Versatil%20testing)%20sundress&fl=name&defType=edismax&mm=2<3 4<5 
6<-80%25&qf=name^10%20partNumber_ntk&debugQuery=true&wt=xml&rows=100

parsedQuery:

+(DisjunctionMaxQuery((((name:versatil name:test)~2)^10.0 | 
partNumber_ntk:versatil testing)) DisjunctionMaxQuery(((name:sundress)^10.0 | 
partNumber_ntk:sundress)))~2

Which seems it incorrect parse name to: name:versatil name:test

If I change the query fields to same fieldType, for example,shortDescription is 
in same fieldType of name:

q=(Versatil%20testing)%20sundress&fl=name&defType=edismax&mm=2<3 4<5 
6<-80%25&qf=name^10%20shortDescription&debugQuery=true&wt=xml&rows=100

ParsedQuery:

+((DisjunctionMaxQuery(((name:versatil)^10.0 | shortDescription:versatil)) 
DisjunctionMaxQuery(((name:test)^10.0 | shortDescription:test))) 
DisjunctionMaxQuery(((name:sundress)^10.0 | shortDescription:sundress)))~2

which hits correctly.

Could someone check this or tell us a quick workaround? Now it have big impact 
on customer.

Thanks in advance! The following is backup information:

 

 

 



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

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

Reply via email to