Weird results with Solr 1.4 and EdgeNGrams - some substrings match, some don't

EDIT 3: The workaround I'm using right now is to strip anything but letters, digits, and whitespace from both my queries and my indexed fields. This produces the desired behavior, but it's very much a workaround rather than a true solution, and I would still like to understand why Solr is doing what it's doing...so still interested in an answer, if anyone has one. END EDIT 3

I have a document named "TT-14B" indexed by Solr 1.4 (via Django/Haystack). When I query the content_auto field for "tt-1" or "tt 14" or "tt 14b" I get the document back; when I query "tt-14" or "tt-14b" I get no results. I edited the Haystack-generated Solr schema a bit to try to fix this, to no avail. Using analyze.jsp, it seems to me that I should be getting a match for "tt-14"; I should certainly be getting one for "tt-14b". (Edit: Oh, and changing the default operator from AND to OR doesn't help.)

Can someone help me understand why this isn't working? Thanks.

...

results

QUERY  | WORKS
=======|======
tt     | yes
tt-    | yes
tt-1   | yes
tt-14  | no
tt-14b | no
tt 14  | yes
tt 14b | yes

EDIT 2

Got some more comparably weird results, might help debug the problem. In this case the test document was "abc'def".

QUERY   | WORKS
========|======
abc     | yes
abc'd   | yes
abc'de  | no
abc'def | no

Same pattern, obviously, but I don't understand what's causing it.

END EDIT 2

schema.xml relevant part (full file below)

<fieldType name="edge_ngram" class="solr.TextField" positionIncrementGap="1">
  <analyzer type="index">
    <tokenizer class="solr.WhitespaceTokenizerFactory" />
    <filter class="solr.LowerCaseFilterFactory" />
    <filter class="solr.WordDelimiterFilterFactory" generateWordParts="1" generateNumberParts="1" splitOnNumerics="0" preserveOriginal="1" catenateWords="0" catenateNumbers="0" catenateAll="0" splitOnCaseChange="1"/>
    <filter class="solr.EdgeNGramFilterFactory" minGramSize="1" maxGramSize="15" side="front" />
  </analyzer>
  <analyzer type="query">
    <tokenizer class="solr.WhitespaceTokenizerFactory" />
    <filter class="solr.LowerCaseFilterFactory" />
    <filter class="solr.WordDelimiterFilterFactory" generateWordParts="1" generateNumberParts="1" splitOnNumerics="0" preserveOriginal="1" catenateWords="0" catenateNumbers="0" catenateAll="0" splitOnCaseChange="1"/>
  </analyzer>
</fieldType>

schema.xml (full)

<?xml version="1.0" ?>
<!--
 Licensed to the Apache Software Foundation (ASF) under one or more
 contributor license agreements.  See the NOTICE file distributed with
 this work for additional information regarding copyright ownership.
 The ASF licenses this file to You under the Apache License, Version 2.0
 (the "License"); you may not use this file except in compliance with
 the License.  You may obtain a copy of the License at

     http://www.apache.org/licenses/LICENSE-2.0

 Unless required by applicable law or agreed to in writing, software
 distributed under the License is distributed on an "AS IS" BASIS,
 WITHOUT WARR开发者_开发知识库ANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License.
-->

<schema name="default" version="1.1">
  <types>
    <fieldtype name="string"  class="solr.StrField" sortMissingLast="true" omitNorms="true"/>
    <fieldType name="boolean" class="solr.BoolField" sortMissingLast="true" omitNorms="true"/>

    <!-- Numeric field types that manipulate the value into
         a string value that isn't human-readable in its internal form,
         but with a lexicographic ordering the same as the numeric ordering,
         so that range queries work correctly. -->
    <fieldType name="sint" class="solr.SortableIntField" sortMissingLast="true" omitNorms="true"/>
    <fieldType name="slong" class="solr.SortableLongField" sortMissingLast="true" omitNorms="true"/>
    <fieldType name="sfloat" class="solr.SortableFloatField" sortMissingLast="true" omitNorms="true"/>
    <fieldType name="sdouble" class="solr.SortableDoubleField" sortMissingLast="true" omitNorms="true"/>

    <fieldType name="date" class="solr.DateField" sortMissingLast="true" omitNorms="true"/>

    <fieldType name="text" class="solr.TextField" positionIncrementGap="100">
      <analyzer type="index">
        <tokenizer class="solr.WhitespaceTokenizerFactory"/>
        <!-- in this example, we will only use synonyms at query time
        <filter class="solr.SynonymFilterFactory" synonyms="index_synonyms.txt" ignoreCase="true" expand="false"/>
        -->
        <filter class="solr.StopFilterFactory" ignoreCase="true" words="stopwords.txt"/>
        <filter class="solr.WordDelimiterFilterFactory" generateWordParts="1" generateNumberParts="1" splitOnNumerics="0" preserveOriginal="1" catenateWords="1" catenateNumbers="1" catenateAll="1"/>
        <filter class="solr.LowerCaseFilterFactory"/>
        <filter class="solr.EnglishPorterFilterFactory" protected="protwords.txt"/>
        <filter class="solr.RemoveDuplicatesTokenFilterFactory"/>
      </analyzer>
      <analyzer type="query">
        <tokenizer class="solr.WhitespaceTokenizerFactory"/>
        <filter class="solr.SynonymFilterFactory" synonyms="synonyms.txt" ignoreCase="true" expand="true"/>
        <filter class="solr.StopFilterFactory" ignoreCase="true" words="stopwords.txt"/>
        <filter class="solr.WordDelimiterFilterFactory" generateWordParts="1" generateNumberParts="1" splitOnNumerics="0" preserveOriginal="1" catenateWords="0" catenateNumbers="0" catenateAll="1"/>
        <filter class="solr.LowerCaseFilterFactory"/>
        <filter class="solr.EnglishPorterFilterFactory" protected="protwords.txt"/>
        <filter class="solr.RemoveDuplicatesTokenFilterFactory"/>
      </analyzer>
    </fieldType>

    <fieldType name="text_ws" class="solr.TextField" positionIncrementGap="100">
      <analyzer>
        <tokenizer class="solr.WhitespaceTokenizerFactory"/>
      </analyzer>
    </fieldType>

    <fieldType name="ngram" class="solr.TextField" >
      <analyzer type="index">
        <tokenizer class="solr.KeywordTokenizerFactory"/>
        <filter class="solr.LowerCaseFilterFactory"/>
        <filter class="solr.NGramFilterFactory" minGramSize="2" maxGramSize="15" />
      </analyzer>
      <analyzer type="query">
        <tokenizer class="solr.KeywordTokenizerFactory"/>
        <filter class="solr.LowerCaseFilterFactory"/>
      </analyzer>
    </fieldType>

    <fieldType name="edge_ngram" class="solr.TextField" positionIncrementGap="1">
      <analyzer type="index">
        <tokenizer class="solr.WhitespaceTokenizerFactory" />
        <filter class="solr.LowerCaseFilterFactory" />
        <filter class="solr.WordDelimiterFilterFactory" generateWordParts="1" generateNumberParts="1" splitOnNumerics="0" preserveOriginal="1" catenateWords="0" catenateNumbers="0" catenateAll="0" splitOnCaseChange="1"/>
        <filter class="solr.EdgeNGramFilterFactory" minGramSize="1" maxGramSize="15" side="front" />
      </analyzer>
      <analyzer type="query">
        <tokenizer class="solr.WhitespaceTokenizerFactory" />
        <filter class="solr.LowerCaseFilterFactory" />
        <filter class="solr.WordDelimiterFilterFactory" generateWordParts="1" generateNumberParts="1" splitOnNumerics="0" preserveOriginal="1" catenateWords="0" catenateNumbers="0" catenateAll="0" splitOnCaseChange="1"/>
      </analyzer>
    </fieldType>
  </types>

  <fields>   
    <!-- general -->
    <field name="id" type="string" indexed="true" stored="true" multiValued="false" required="true"/>
    <field name="django_ct" type="string" indexed="true" stored="true" multiValued="false" />
    <field name="django_id" type="string" indexed="true" stored="true" multiValued="false" />

    <dynamicField name="*_i"  type="sint"    indexed="true"  stored="true"/>
    <dynamicField name="*_s"  type="string"  indexed="true"  stored="true"/>
    <dynamicField name="*_l"  type="slong"   indexed="true"  stored="true"/>
    <dynamicField name="*_t"  type="text"    indexed="true"  stored="true"/>
    <dynamicField name="*_b"  type="boolean" indexed="true"  stored="true"/>
    <dynamicField name="*_f"  type="sfloat"  indexed="true"  stored="true"/>
    <dynamicField name="*_d"  type="sdouble" indexed="true"  stored="true"/>
    <dynamicField name="*_dt" type="date"    indexed="true"  stored="true"/>


    <field name="modelname_exact" type="string" indexed="true" stored="true" multiValued="false" />

    <field name="modelname" type="text" indexed="true" stored="true" multiValued="false" />

    <field name="name" type="text" indexed="true" stored="true" multiValued="false" />

    <field name="text" type="text" indexed="true" stored="true" multiValued="false" />

    <field name="name_exact" type="string" indexed="true" stored="true" multiValued="false" />

    <field name="content_auto" type="edge_ngram" indexed="true" stored="true" multiValued="true" />

  </fields>

  <!-- field to use to determine and enforce document uniqueness. -->
  <uniqueKey>id</uniqueKey>

  <!-- field for the QueryParser to use when an explicit fieldname is absent -->
  <defaultSearchField>text</defaultSearchField>

  <!-- SolrQueryParser configuration: defaultOperator="AND|OR" -->
  <solrQueryParser defaultOperator="AND" />
</schema>

An screenshout of the /admin/analysis.jsp for every case would be interesting.

Is there a reason, why positionIncrementGap="1"is set to 1?

tt-14b and tt 14b are handled different, because of the whitspace tokenizer. That means: tt-14b is one term, as long as the WordDelimiterFilterFactorydoesn't fired, while tt 14b are 2 terms from the beginning. The positionIncrementGap gives you the possibility to see different terms as one phrase, even if there are not neighbors, but on the next "n" position. So try to rise the positionIncrementGap.

Btw: The first i notice on your schema.xml are the missing "EdgeNGramFilterFactory" at query time. Which should be okay. But there are also understandable reasons, why "same filters on query- and index-time" are handled as best practices. This depends on every special situation, but activating this filter on query-time would be a try.

A little late to the show on this one -- but, as noted above, the WhitespaceTokenizerFactory breaks words with hyphens if it's passed through the StandardAnalyzer. I found this out the hard way too...:

Solr NGramTokenizerFactory and PatternReplaceCharFilterFactory - Analyzer results inconsistent with Query Results

The solution is probably to use the KeywordAnalyzer -- it shouldn't split anything.

I came up with a similar work-around to your "EDIT 3" on the link above (in PHP).

The particularly frustrating thing about the Solr analyzer is that it shows everything is fine and behaving as expected -- which really confused me.

Good luck!