Sunspot - Solr Admin returns results for "-1" but Sunspot Gem does not

Question

I have Sunspot/Solr deployed on a site for searching. I want partial searching on Part Numbers with a hyphen down to two characters.

My current versions are:

Solr & Lucense 3.5
sunspot (2.0.0)
sunspot_rails (2.0.0)
sunspot_solr (2.0.0)

My config file:

<fieldType name="n_gram_text" class="solr.TextField" omitNorms="false">
  <analyzer type="index">
    <tokenizer class="solr.StandardTokenizerFactory"/>
    <filter class="solr.StandardFilterFactory"/>
    <filter class="solr.NGramFilterFactory" minGramSize="2" maxGramSize="15"/>
  </analyzer>
  <analyzer type="query">
    <tokenizer class="solr.WhitespaceTokenizerFactory"/>
    <filter class="solr.WordDelimiterFilterFactory" preserveOriginal="1"/>
  </analyzer>
</fieldType>

<dynamicField name="*_ngram" stored="false" type="n_gram_text" multiValued="true" indexed="true"/>

Sample Part Number:

455880-1

So if I use the SOLR Admin console and search for "-1" I get results back. In the analyzer I confirmed that "-1" was a gram and that the search query "-1" matches it and "0-1".

But when I perform this search in my website it fails. If I search "0-1" I'll get results but if it's just "-1" it give me nothing. I've tried escaping it with "-1" but that doesn't change the outcome.

What else could I troubleshoot between the working Solr and not working Sunspot?

In my Rails logs I have the following:

SOLR Request (7.5ms)  [ path=#<RSolr::Client:0x000001081efe70> 
parameters={data: 
  fq=type%3AGroup
  &fq=is_site_b%3Atrue
  &q=-1
  &fl=%2A+score
  &qf=name_text+display_ngram
  &defType=dismax
  &start=0
  &rows=20, 
method: post, 
params: {:wt=>:ruby}, 
query: wt=ruby, 
headers: {"Content-Type"=>"application/x-www-form-urlencoded; charset=UTF-8"}, path: select, uri: http://localhost:8984/solr/select?wt=ruby, open_timeout: , read_timeout: , retry_503: , retry_after_limit: } ]

Model Setup:

searchable do
  text     :name
  text     :item_display_part_numbers, :as  => :display_ngram
end

def item_display_part_numbers
  self.items.map(&:display_part_number)
end

item_display_part_numbers is an array of part numbers. The patterns are either digits, digits with a -1, or the text "n/a".

Search :

@search = Sunspot.search(Group) do
   fulltext params[:search_string]
   paginate(:page => params[:page], :per_page => params[:per_page] || 20)
end

I believe this data object is indexed correctly. In the console if I retrieve it and call it's Index method I get the following:

<?xml version="1.0" encoding="UTF-8"?>
<add>
   <doc>
      <field name="id">Group 1365</field>
      <field name="type">Group</field>
      <field name="type">ActiveRecord::Base</field>
      <field name="class_name">Group</field>
      <field name="name_text">HEAVY DUTY BRASS BELL</field>
      <field name="display_ngram">n/a</field>
      <field name="display_ngram">455880-1</field>
      <field name="display_ngram">n/a</field>
   </doc>
</add>

Answer 1

You are using the DismaxQueryParser. If a query starts with a hyphen, the following text is interpreted as prohibited (http://wiki.apache.org/solr/DisMaxQParserPlugin#Query_Syntax).
DismaxQueryParser supports phrase searching.

A solution could be to adjust the solr params:

 @search = Sunspot.search(Group) do
   adjust_solr_params do |params|
    params[:q] = "\"#{params[:q]}\"" if params[:q].start_with?("-")
   end      
   fulltext params[:search_string]
   paginate(:page => params[:page], :per_page => params[:per_page] || 20)
 end

It looks like a dirty hack - but it works...