Problems with solr docker suggester in context query: blanks are not correctly processed

Question

I am using Solr version 8.8.1 in a docker container. I use an address query system based on a CSV file. I want to get a city in relation to the street name. The fields in my managed-schema are looking like this:

<fieldType name="text_suggest" class="solr.TextField" positionIncrementGap="100">
  <analyzer>
   <tokenizer class="solr.KeywordTokenizerFactory"/>
            <filter class="solr.LowerCaseFilterFactory"/>
   </analyzer>
</fieldType>

<field name="sugAddress" type="text_suggest" multiValued="true" indexed="true" stored="true"/>
<field name="sugCity" type="text_suggest"  indexed="true" stored="true"/>
 <copyField source="CITY" dest="sugAddress"/>
 <copyField source="POSTCODE" dest="sugAddress"/>
 <copyField source="STREET" dest="sugAddress"/>
 <copyField source="CITY" dest="sugCity"/>

My solrconfig.xml is looking like this:

<searchComponent name="solr_suggest" class="solr.SuggestComponent">
    <lst name="suggester">
    <str name="name">sugCitySuggester</str>
    <str name="lookupImpl">BlendedInfixLookupFactory</str>
    <str name="highlight">false</str>
   <str name="dictionaryImpl">DocumentDictionaryFactory</str> 
    <str name="field">sugCity</str>
    <str name="buildOnOptimize">true</str>
    <str name="weightField">sugAddress</str> 
    <str name="contextField">sugAddress</str>
    <str name="suggestAnalyzerFieldType">text_suggest</str>
    <str name="buildOnCommit">true</str>
  </lst>
</searchComponent>

<requestHandler name="/suggest" class="solr.SearchHandler" startup="lazy">
  <lst name="defaults">
      <str name="suggest">true</str>
      <str name="suggest.count">10</str>
  </lst>
  <arr name="components">
    <str>solr_suggest</str>
  </arr>
</requestHandler>

The result of a normal query (without a suggester) looks like this:

{
 "sugAddress":["Am Bahndamm",
          "3",
          "3",
          "65432",
          "Testcity"],
        "sugStreet":"Am Bahndamm",
        "sugPostcode":"65432",
        "sugCity":"Testcity",
        "id":"1e470f44-fe3a-4fb9-b71d-ca556a48c55f",
        "_version_":1695636026132791297,
        "NO_FROM":[3],
        "NO_TO":[3],
        "POSTCODE":[65432]
},
{
 "sugAddress":["Teststreet",
          "4",
          "4",
          "65432",
          "Testcity"],
        "sugStreet":"Teststreet",
        "sugPostcode":"65432",
        "sugCity":"Testcity",
        "id":"1e470f44-fe3a-4fb9-b71d-ca556a48c55f",
        "_version_":1695636026132791297,
        "NO_FROM":[4],
        "NO_TO":[4],
        "POSTCODE":[65432]
}

All indexes are build.

The following query brings correctly one result:

http://localhost:8983/solr/addrescore/suggest?suggest=true&suggest.dictionary=sugCitySuggester&suggest.q=Test&suggest.cfq=Teststreet

But the following returns no results:

http://localhost:8983/solr/addrescore/suggest?suggest=true&suggest.dictionary=sugCitySuggester&suggest.q=Test&suggest.cfq=Am%20Bahndamm

There seems to be a problem with spaces in the context query. Has someone had an idea?

Answer 1

Suggester works on terms. That is why, first of all, open Solr admin console and check terms for sugStreet field. According to your question, this information should be availably by this URL: http://localhost:8983/solr/#/addrescore/schema?field=sugStreet

I expect that "Am Bahnda" will not be on that list. Then your goal will be to define, why this street is not on the list of terms. I suspect that it can be because of how sugStreet field is configured. In order to have "Am Bahndamm" as one term, you need to make it untokenized(by default all Sitecore search fields are tokenized). Tokenized field means that your string is split by dividers(spaces, tabs, commas). I expect that "Am Bahndamm" was divided into 2 terms: "Am" and "Bahndamm". And "Am" was ignored due to length shorter than the 3 symbols.

You will need to change sugStreet field configuration and make it untokenized.

indexType="UNTOKENIZED"

Then rebuild indexes and check your terms on Solr admin console to make sure that it started to work.

UPDATE: As far as I understand, you want to achieve it without touching Sitecore configuration.

I this case you need to define your sugStreet field with new type without tokenizer by spaces:

<field name="sugStreet" type="text_suggest" indexed="true" stored="true"/>

Then you need to copy source field to destination:

<copyField source="STREET" dest="sugStreet"/>

Afterward, you need to rebuild the index. Copying of field value happens after analisis, so it should work. You should be able to get untokenized terms in sugStreet field and tokenized in STREET field.