Elasticsaerch Query Exact match

[Mohan_T]

to find the exact match

mappings

"keywords": {
    "type": "object",
    "enabled": True,
    "properties": {
        "keyword_values": {
            "type": "text",
            "analyzer": "synonym_stemmer_bad_words_analyzer",
            "fields": {
                "keyword": {
                    "type": "keyword"
                }
            }
        }
    }
}

Doc 1

"keywords": [
    {
        "keyword_values": "marbles"
    },
    {
        "keyword_values": "flooring"
    }
}

Doc 2

"keywords": [
    {
        "keyword_values": "marble"
    },
    {
        "keyword_values": "floor"
    }
}

Doc 3

"keywords": [
    {
        "keyword_values": "marble floor"
    },
    {
        "keyword_values": "floor marble"
    }
}

If am searching for marble with keywords

using the below query

Query: "keywords.keyword_values"

{
    "query": {
        "bool": {
            "should": [
                {
                    "match": {
                        "keywords.keyword_values": {
                            "query": "marble",
                            "operator": "AND",
                            "prefix_length": 0,
                            "max_expansions": 50,
                            "fuzzy_transpositions": "true",
                            "lenient": "false",
                            "zero_terms_query": "NONE",
                            "auto_generate_synonyms_phrase_query": "true",
                            "boost": 7
                        }
                    }
                }
            ],
            "minimum_should_match": "1"
        }
    }
}

I will be getting all three doc results, but I need to fetch only Doc 1 & Doc 2

so I have used the query with the keyword

Query: "keywords.keyword_values.keyword"

{
    "query": {
        "bool": {
            "should": [
                {
                    "match": {
                        "keywords.keyword_values.keyword": {
                            "query": "marble",
                            "operator": "AND",
                            "prefix_length": 0,
                            "max_expansions": 50,
                            "fuzzy_transpositions": "true",
                            "lenient": "false",
                            "zero_terms_query": "NONE",
                            "auto_generate_synonyms_phrase_query": "true",
                            "boost": 7
                        }
                    }
                }
            ],
            "minimum_should_match": "1"
        }
    }
}

in this am getting only Doc 2, here i need to fetch Doc 1 & Doc 2

[RabBit_BR (andre.coelho)]

Hi @Mohan_T

Please provide the analyzer below:

@Mohan_T

synonym_stemmer_bad_words_analyzer

[Mohan_T]

{
    "tokens": [
        {
            "token": "marbl",
            "start_offset": 0,
            "end_offset": 6,
            "type": "<ALPHANUM>",
            "position": 0
        }
    ]
}

[Mohan_T]

@RabBit_BR I have updated the analyzer, and the above one am getting the results when am using analyzer

[RabBit_BR (andre.coelho)]

I don't know if you can create a new field and new analyzer but in the example below it was possible to retrieve the docs using the keyword tokenizer.

PUT idx_test
{
  "settings": {
    "analysis": {
      "analyzer": {
        "my_analyzer": {
          "tokenizer": "keyword",
          "filter": [
            "my_filter"
          ]
        }
      },
      "filter": {
        "my_filter": {
          "type": "stemmer"
        }
      }
    }
  },
  "mappings": {
    "properties": {
      "keywords": {
        "properties": {
          "keyword_values": {
            "type": "text",
            "analyzer": "my_analyzer",
            "fields": {
              "keyword": {
                "type": "keyword"
              }
            }
          }
        }
      }
    }
  }
}

Query

GET idx_test/_search
{
  "query": {
    "match": {
      "keywords.keyword_values": "marbles"
    }
  }
}

[Mohan_T]

@RabBit_BR Thank you, it's working as expected, but I have one more use case, for example, Docs 4 & 5 has the below data.

Doc 4

"keywords": [
    {
        "keyword_values": "marble floor"
    }
}

Doc 5

"keywords": [
    {
        "keyword_values": "floor marble"
    }
}

Search Query

{
    "query": {
        "bool": {
            "should": [
                {
                    "match": {
                        "keywords.keyword_values": {
                            "query": "marble floor",
                            "operator": "AND",
                            "prefix_length": 0,
                            "max_expansions": 50,
                            "fuzzy_transpositions": "true",
                            "lenient": "false",
                            "zero_terms_query": "NONE",
                            "auto_generate_synonyms_phrase_query": "true",
                            "boost": 7
                        }
                    }
                }
            ],
            "minimum_should_match": "1"
        }
    }
}

I want to get Docs 4 & 5 but am getting only Doc 4 using the above query

[RabBit_BR (andre.coelho)]

I saw and other post similiar question and the orientation was using span fist query, maybe work for you.