CODEWITHSUNDEEP
elasticsearchelasticsearch-nested
SSG
SSG

Reputation: 1505

Elasticsearch : minimum_should_match for nested query

I have a nested field like 

{
  "tags": [
    {
      "tag": "lorem ipsum"
    },
    {
      "tag": "Lorem ipsum dolor sit amet"
    }
  ]
}

And mapping like

{
  "tags": {
    **"type": "nested",**
    "properties": {
      "tag": {
        "type": "text",
        "fields": {
          "keyword": {
            "type": "keyword",
            "ignore_above": 256
          }
        }
      }
    }
  }
}

can we use something like minimum_should_match : 80 for a nested tag field? So that I will be able to control the relevance level through it?
Exa: if I search for "Lorem ipsum dolor" with minimum_should_match: 90 , I should not get lorem ipsum as a result.

Upvotes: 0

Views: 921

Answers (1)

jaspreet chahal
jaspreet chahal

Reputation: 9099

Nested query is just a syntax to access nested fields so minimum_should_match can be used as in other queries

Query

{
  "query": {
    "nested": {
      "path": "tags",
      "query": {
        "match": {
          "tags.tag": 
          {
            "query": "lorem ipsum dolor",
            "minimum_should_match": "90%"
          }
        }
      },
      "inner_hits": {}
    }
  }
}

Result:

  "hits" : {
    "total" : {
      "value" : 1,
      "relation" : "eq"
    },
    "max_score" : 0.671082,
    "hits" : [
      {
        "_index" : "index56",
        "_type" : "_doc",
        "_id" : "01We63ABq1Ib1oOmkJxn",
        "_score" : 0.671082,
        "_source" : {
          "tags" : [
            {
              "tag" : "lorem ipsum"
            },
            {
              "tag" : "Lorem ipsum dolor sit amet"
            }
          ]
        },
        "inner_hits" : {
          "tags" : {
            "hits" : {
              "total" : {
                "value" : 2,
                "relation" : "eq"
              },
              "max_score" : 0.89999837,
              "hits" : [
                {
                  "_index" : "index56",
                  "_type" : "_doc",
                  "_id" : "01We63ABq1Ib1oOmkJxn",
                  "_nested" : {
                    "field" : "tags",
                    "offset" : 1
                  },
                  "_score" : 0.89999837,
                  "_source" : {
                    "tag" : "Lorem ipsum dolor sit amet"
                  }
                },
                {
                  "_index" : "index56",
                  "_type" : "_doc",
                  "_id" : "01We63ABq1Ib1oOmkJxn",
                  "_nested" : {
                    "field" : "tags",
                    "offset" : 0
                  },
                  "_score" : 0.44216567,
                  "_source" : {
                    "tag" : "lorem ipsum"
                  }
                }
              ]
            }
          }
        }
      }
    ]
  }

With minimum-should-match:90% both the nested docs are returned in inner_hits. Reason: From docs

The number computed from the percentage is rounded down and used as the minimum.

Since 90% of 22.7 it will be rounded down to 2. So 2 tokens should match. If minimum-should-match:100% then only one nested doc will be returned

Upvotes: 1

Related Questions