Question

I’m trying to setup the mapping for my elasticsearch instance with full name matching and partial name matching:

curl -XPUT 'http://127.0.0.1:9200/test/?pretty=1'  -d '{
  "mappings": {
    "venue": {
      "properties": {
        "location": {
          "type": "geo_point"
        },
        "name": {
          "fields": {
            "name": {
              "type": "string",
              "analyzer": "full_name"
            },
            "partial": {
              "search_analyzer": "full_name",
              "index_analyzer": "partial_name",
              "type": "string"
            }
          },
          "type": "multi_field"
        }
      }
    }
  },
  "settings": {
    "analysis": {
      "filter": {
        "swedish_snow": {
          "type": "snowball",
          "language": "Swedish"
        },
        "name_synonyms": {
          "type": "synonym",
          "synonyms_path": "name_synonyms.txt"
        },
        "name_ngrams": {
          "side": "front",
          "min_gram": 2,
          "max_gram": 50,
          "type": "edgeNGram"
        }
      },
      "analyzer": {
        "full_name": {
          "filter": [
            "standard",
            "lowercase"
          ],
          "type": "custom",
          "tokenizer": "standard"
        },
        "partial_name": {
          "filter": [
            "swedish_snow",
            "lowercase",
            "name_synonyms",
            "name_ngrams",
            "standard"
          ],
          "type": "custom",
          "tokenizer": "standard"
        }
      }
    }
  }
}'

I fill it with some data:

curl -XPOST 'http://127.0.0.1:9200/_bulk?pretty=1'  -d '
{"index" : {"_index" : "test", "_type" : "venue"}}
{"location" : [59.3366, 18.0315], "name" : "johnssons"}
{"index" : {"_index" : "test", "_type" : "venue"}}
{"location" : [59.3366, 18.0315], "name" : "johnsson"}
{"index" : {"_index" : "test", "_type" : "venue"}}
{"location" : [59.3366, 18.0315], "name" : "jöhnsson"}
'

Perform some searches to test,
Full name:

curl -XGET 'http://127.0.0.1:9200/test/venue/_search?pretty=1' -d '{
  "query": {
    "bool": {
      "should": [
        {
          "text": {
            "name": {
              "boost": 1,
              "query": "johnsson"
            }
          }
        },
        {
          "text": {
            "name.partial": "johnsson"
          }
        }
      ]
    }
  }
}'

Result:

{
  "took": 3,
  "timed_out": false,
  "_shards": {
    "total": 5,
    "successful": 5,
    "failed": 0
  },
  "hits": {
    "total": 2,
    "max_score": 0.29834434,
    "hits": [
      {
        "_index": "test",
        "_type": "venue",
        "_id": "CAO-dDr2TFOuCM4pFfNDSw",
        "_score": 0.29834434,
        "_source": {
          "location": [
            59.3366,
            18.0315
          ],
          "name": "johnsson"
        }
      },
      {
        "_index": "test",
        "_type": "venue",
        "_id": "UQWGn8L9Squ5RYDMd4jqKA",
        "_score": 0.14663845,
        "_source": {
          "location": [
            59.3366,
            18.0315
          ],
          "name": "johnssons"
        }
      }
    ]
  }
}

Partial name:

curl -XGET 'http://127.0.0.1:9200/test/venue/_search?pretty=1' -d '{
  "query": {
    "bool": {
      "should": [
        {
          "text": {
            "name": {
              "boost": 1,
              "query": "johns"
            }
          }
        },
        {
          "text": {
            "name.partial": "johns"
          }
        }
      ]
    }
  }
}'

Result:

{
  "took": 3,
  "timed_out": false,
  "_shards": {
    "total": 5,
    "successful": 5,
    "failed": 0
  },
  "hits": {
    "total": 2,
    "max_score": 0.14663845,
    "hits": [
      {
        "_index": "test",
        "_type": "venue",
        "_id": "UQWGn8L9Squ5RYDMd4jqKA",
        "_score": 0.14663845,
        "_source": {
          "location": [
            59.3366,
            18.0315
          ],
          "name": "johnssons"
        }
      },
      {
        "_index": "test",
        "_type": "venue",
        "_id": "CAO-dDr2TFOuCM4pFfNDSw",
        "_score": 0.016878016,
        "_source": {
          "location": [
            59.3366,
            18.0315
          ],
          "name": "johnsson"
        }
      }
    ]
  }
}

Name within name:

curl -XGET 'http://127.0.0.1:9200/test/venue/_search?pretty=1' -d '{
  "query": {
    "bool": {
      "should": [
        {
          "text": {
            "ame": {
              "boost": 1,
              "query": "johnssons"
            }
          }
        },
        {
          "text": {
            "name.partial": "johnssons"
          }
        }
      ]
    }
  }
}'

Result:

{
  "took": 2,
  "timed_out": false,
  "_shards": {
    "total": 5,
    "successful": 5,
    "failed": 0
  },
  "hits": {
    "total": 1,
    "max_score": 0.39103588,
    "hits": [
      {
        "_index": "test",
        "_type": "venue",
        "_id": "UQWGn8L9Squ5RYDMd4jqKA",
        "_score": 0.39103588,
        "_source": {
          "location": [
            59.3366,
            18.0315
          ],
          "name": "johnssons"
        }
      }
    ]
  }
}

As you can see I’m only getting one venue back which is johnssons. Shouldn’t I get both johnssons and johnsson back? What am I doing wrong in my settings?

Answer 1

You are using full_name analyzed as a search analyzer for the name.partial field. As a result your query is getting translated into the query for the term johnssons, which doesn’t match anything.

You can use Analyze API to see what how your records are indexed. For example, this command

curl -XGET 'http://127.0.0.1:9200/test/_analyze?analyzer=partial_name&pretty=1' -d 'johnssons'

will show you that during indexing the string “johnssons” is getting translated into the following terms: “jo”, “joh”, “john”, “johns”, “johnss”, “johnsso”, “johnsson”. While this command

 curl -XGET 'http://127.0.0.1:9200/test/_analyze?analyzer=full_name&pretty=1' -d 'johnssons'

will show you that during searching the string “johnssons” is getting translated into term “johnssons”. As you can see there is no match between your search term and your data here.