Elasticsearch - find items in array with score

Question

I need to return a scored list of posts that contain as many tags that user provided as possible. Tags are stored in arrays as objects and user request contains an array of tag IDs (eg. ["a", "c"]). Single document looks something like this:

{
  id: 1,
  meta: {
    tags: [{id: "a", name: "Engine"}, {id: "b", name: "Street"}, {id: "c", name: "Sport"}]
  }
}

Sadly, term query returns posts in almost random order, as all of them have score 1.0:

{
  "query": {
    "bool": {
      "must": [
        {
          "terms": {
            "meta.tags.id.keyword": [
             "a", "c"
            ]
          }
        }
      ]
    }
  }
}

I assume that thing would be quite easy if all tags would be in a single string, as this would be normal full text search, but how to achieve something like this for object arrays?

Here is a mapping exception (dynamically created by NEST client):

"mappings" : {
  "properties" : {
    "categories" : {
      "properties" : {
        "id" : {
          "type" : "text",
          "fields" : {
            "keyword" : {
              "type" : "keyword",
              "ignore_above" : 256
            }
          }
        },
        "name" : {
          "type" : "text",
          "fields" : {
            "keyword" : {
              "type" : "keyword",
              "ignore_above" : 256
            }
          }
        }
      }
    },
    "description" : {
      "type" : "text",
      "fields" : {
        "keyword" : {
          "type" : "keyword",
          "ignore_above" : 256
        }
      }
    },
    "id" : {
      "type" : "text",
      "fields" : {
        "keyword" : {
          "type" : "keyword",
          "ignore_above" : 256
        }
      }
    },
    "inFavourite" : {
      "type" : "long"
    },
    "likes" : {
      "type" : "long"
    },
    "name" : {
      "type" : "text",
      "fields" : {
        "keyword" : {
          "type" : "keyword",
          "ignore_above" : 256
        }
      }
    },
    "photoPath" : {
      "type" : "text",
      "fields" : {
        "keyword" : {
          "type" : "keyword",
          "ignore_above" : 256
        }
      }
    },
    "meta" : {
      "properties" : {
        "tags" : {
          "properties" : {
            "id" : {
              "type" : "text",
              "fields" : {
                "keyword" : {
                  "type" : "keyword",
                  "ignore_above" : 256
                }
              }
            },
            "name" : {
              "type" : "text",
              "fields" : {
                "keyword" : {
                  "type" : "keyword",
                  "ignore_above" : 256
                }
              }
            },
          }
        },
      }
    },
    "userId" : {
      "type" : "text",
      "fields" : {
        "keyword" : {
          "type" : "keyword",
          "ignore_above" : 256
        }
      }
    },
    "viewCounter" : {
      "type" : "long"
    }
  }
}

Good question, would you mind sharing your index mapping, few sample and expected docs, so that its easy for community to reproduce your issue and provide working solution on your data-set. — Amit
– Amit, Commented Feb 7, 2021 at 3:50
I hava added a mapping. Docs are quite big, so it would be probably better to work on some custom examples. — Janusz Guzowski
– Janusz Guzowski, Commented Feb 7, 2021 at 9:01

KennetsuR · Accepted Answer · 2021-02-07 09:47:45Z

1

you could like to boost different query score, like the question Elasticsearch sort based on the number of occurrences a string appears in an array
you also would like to change the logic relationship of conditions.e.g. cond1 or cond2 or cond3 -> (cond1 & cond2) | cond3.
and then you might like to write custom score function to calculate e.g.

{
  "query": {
    "function_score": {
      "query": { "match_all": {} },
      "boost": "5", 
      "functions": [
        {
          "filter": { "match": { "test": "bar" } },
          "random_score": {}, 
          "weight": 23
        },
        {
          "filter": { "match": { "test": "cat" } },
          "weight": 42
        }
      ],
      "max_boost": 42,
      "score_mode": "max",
      "boost_mode": "multiply",
      "min_score": 42
    }
  }
}

answered Feb 7, 2021 at 9:47

KennetsuR

78610 silver badges20 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Elasticsearch - find items in array with score

1 Answer 1

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related