Elasticsearch: Query avoiding defined analyzer / using another analyzer

1
votes

I have a question regarding to ElasticSearch queries. I have a field called “name” that is defined with an analyzer from the phonetic plugin.

  "name" : {
      "type" : "string",
      "store" : "yes",
      "index" : "analyzed",
      "analyzer" : "soundsAnalyzer",
      "boost" : 10.0
  },

Everything works fine so far! However I have a certain query on this field where I wanna get exact matches rather than similar sounded matches. I tried to accomplish this by using a term query as well as a text query. The term query doesn’t return any result at all and I have no glue why. By using the text query and defining another analyzer I get also nothing. I've tried the following query:

{
  "query": {
    "text": {
      "documents.name": {
        "query": "England",
        "analyzer": "plainAnalyzer"
      }
    }
  }
}

So my question, how can I accomplish avoiding the defined analyzer in the mapping just for this certain query to get exact matches? TIA

1
elasticsearch

1 Answers

1
votes

You will have to index this field twice - one time with soundsAnalyzer and another using plainAnalyzer. You can achieve this using multi_field type:

"name" : {
    "type" : "multi_field",
    "fields" : {
        "name" : {"type" : "string", "store" : "yes", "index" : "analyzed", "analyzer" : "soundsAnalyzer", "boost" : 10.0},
        "plain" : {"type" : "string", "index" : "analyzed", "analyzer" : "plainAnalyzer"}
    }
}

All current queries that are using soundsAnalyzer will remain the same. In queries that require plainAnalyzer you can use name.plain instead of name:

{
  "query": {
    "text": {
      "documents.name.plain": {
        "query": "England"
      }
    }
  }
}