Re: Confusing SOLR results after upgrading from 4.10 to 7.1

[Susheel Kumar]

This may not be the reason but i noticed you have FlattenGraphFilterFactory
at query time while its only required at index time. I would suggest to go
Analysis tab if not checked already.

Thnx


On Mon, Apr 30, 2018 at 2:22 PM, Hodder, Rick  wrote:

> I upgraded from SOLR 4.10 to SOLR 7.1
>
> In the core, I have a string field called "company" and string field
> "year", and I have an index on company called IDX_Company.
> Here is the definition of the company field, and the definition of
> text_general in my schema in 4.10
>
>  stored="false" multiValued="true" />
> 
> 
>  positionIncrementGap="100">
> 
> 
>  class="solr.EdgeNGramFilterFactory"
> minGramSize="1" maxGramSize="15" side="front"/>
>  class="solr.StopFilterFactory" ignoreCase="true" words="stopwords.txt" />
>  class="solr.SynonymFilterFactory"
> synonyms="synonyms.txt" ignoreCase="true" expand="true"/>
> 
> 
>  pattern="[^\p{L}\p{N} ]" replacement=" " replace="all" />
> 
> 
> 
>  class="solr.EdgeNGramFilterFactory"
> minGramSize="1" maxGramSize="15" side="front"/>
>  class="solr.StopFilterFactory" ignoreCase="true" words="stopwords.txt" />
>  class="solr.SynonymFilterFactory"
> synonyms="synonyms.txt" ignoreCase="true" expand="true"/>
> 
>
>   
>  pattern="[^\p{L}\p{N} ]" replacement=" " replace="all" />
> 
> 
>
>
> Here is the field definition and definition of text_general in 7.10
>
>  stored="false" multiValued="true" />
> 
> 
>  positionIncrementGap="100" multiValued="true">
> 
> 
>  maxGramSize="15"/>
>  words="stopwords.txt" />
>  synonyms="synonyms.txt" ignoreCase="true" expand="false"/>
> 
> 
> 
>  pattern="[^\p{L}\p{N} ]" replacement=" " replace="all" />
> 
> 
> 
>  maxGramSize="15"/>
>  words="stopwords.txt" />
>  synonyms="synonyms.txt" ignoreCase="true" expand="false"/> class="solr.FlattenGraphFilterFactory"/>
> 
> 
>  pattern="[^\p{L}\p{N} ]" replacement=" " replace="all" />
> 
> 
>
> Among the documents in the core are:
>
> company year
> AB Landscaping Inc.  2001
> AB Landscaping Inc.  2002
> AB Landscaping : AB Landscaping and Excavating LLC  2001
> AB Landscaping : AB Landscaping and Excavating LLC  2002
> A B Landscaping : AB Landscaping and Excavating LLC  2000
> Landscaping Firm1999
> Landscaping Associates  1998
>
> Under 4.10 if I search for
> IDX_Company:(AB AND Landscaping)
> I see all 7 companies, and the two AB Landscaping Incs are at the top of
> the results
>
> Under 7.1 if I search for
> IDX_Company:(AB AND Landscaping)
>
> I only see the following, notice that documents with Excavating dont
> appear, and AB Landscapting are not at the top of the results - the
> Landscaping Firm and Landscaping Associates are
>
> Landscaping Associates  1998
> Landscaping Firm1999
> AB Landscaping Inc.  2001
> AB Landscaping Inc.  2002
>
> Any ideas that might be causing this? The query seems very straightforward.
>
>

Confusing SOLR results after upgrading from 4.10 to 7.1

[Hodder, Rick]

I upgraded from SOLR 4.10 to SOLR 7.1

In the core, I have a string field called "company" and string field "year", 
and I have an index on company called IDX_Company.
Here is the definition of the company field, and the definition of text_general 
in my schema in 4.10



























Here is the field definition and definition of text_general in 7.10


























Among the documents in the core are:

company year
AB Landscaping Inc.  2001
AB Landscaping Inc.  2002
AB Landscaping : AB Landscaping and Excavating LLC  2001
AB Landscaping : AB Landscaping and Excavating LLC  2002
A B Landscaping : AB Landscaping and Excavating LLC  2000
Landscaping Firm1999
Landscaping Associates  1998

Under 4.10 if I search for
IDX_Company:(AB AND Landscaping)
I see all 7 companies, and the two AB Landscaping Incs are at the top of the 
results

Under 7.1 if I search for
IDX_Company:(AB AND Landscaping)

I only see the following, notice that documents with Excavating dont appear, 
and AB Landscapting are not at the top of the results - the Landscaping Firm 
and Landscaping Associates are

Landscaping Associates  1998
Landscaping Firm1999
AB Landscaping Inc.  2001
AB Landscaping Inc.  2002

Any ideas that might be causing this? The query seems very straightforward.