Elasticsearch 查询权重（qbit）

前言

• 本文对 Elasticsearch 7.17 适用

提升字段的权重

multi_match

• 默认 type 为 best_fields

  GET /_search
  {
  "query": {
    "multi_match": {
      "query": "this is a test",
      "fields": ["subject^3", "message"]
    }
  }
  }

• most_fields

  GET /_search
  {
  "query": {
    "multi_match": {
      "query": "this is a test",
      "type": "most_fields",
      "fields": ["subject^3", "message"]
    }
  }
  }
  // 等价于
  GET /_search
  {
  "query": {
    "bool": {
      "should": [
        {
          "match": { "subject": { "query": "this is a test", "boost": 3 }}
        },
        {
          "macth": { "message": "this is a test" }
        }
      ]
    }
  }
  }

query_string

GET /_search
{
  "query": {
    "query_string": {
      "fields": ["content", "name^5"],
      "query": "this AND that OR thus"
    }
  }
}

simple_query_string

GET /_search
{
  "query": {
    "simple_query_string": {
      "query": """"fried eggs" +(eggplant | potato) -frittata""",
      "fields": ["title^5", "body"],
      "default_operator": "and"
    }
  }
}

提升子句的权重

• bool 查询

  GET /_search
  {
  "query": {
    "bool": {
      "should": [
        {
          "match": { "subject": { "query": "this is a test", "boost": 3 }}
        },
        {
          "macth": { "message": "this is a test" }
        }
      ]
    }
  }
  }

提升索引的权重

• indices_boost

  GET /_search
  {
    "indices_boost" : [
        { "alias1" : 1.4 },
        { "index*" : 1.3 }
    ]
  }

Function score query

• ElasticSearch 相关性打分机制

• 年份越大，分值越高

  GET /_search
  {
  "query": {
    "function_score": {
      "field_value_factor": {
        "field": "year",
        "factor": 1.2,
        "modifier": "sqrt",
        "missing": 1
      }
    }
  }
  }

• 提升英文文档权重（固定字段值）

  GET /_search
  {
  "query": {
    "function_score": {
      "query": {
        "match": {
          "title": "data"
        }
      },
      "script_score": {
        "script": "return doc['language'].value == 'EN' ? 1.5 : 1.0"
      }
    }
  }
  }

• 提升英文文档权重（字段值传参）

  GET zt_title_info_alias/_search
  {
  "query": {
    "function_score": {
      "query": {
        "match": {
          "title": "data"
        }
      },
      "script_score": {
        "script": {
          "params": {
            "recommend_language": "EN"
          },
          "source": "return doc['language'].value == params.recommend_language ? 1.5 : 1.0"
        }
      }
    }
  }
  }

Script score query

• 注意 function_score 子查询的 script_score 不同于一级查询中的 script_score，前者的返回值乘以原 score 为最终 score，后者的返回值即为最终 score

• 示例，给 score 加 10 分

  {
  "query": {
    "script_score": {
      "query": {
        "match": {
          "title": "data"
        }
      },
      "script": {
        "lang": "painless",
        "source": "_score + 10;"
      }
    }
  }
  }

本文出自 qbit snap