Elasticsearch多条件查询及可选匹配排序需求

索引映射（修正后） #

原映射中hobbiles定义为keyword类型，但实际存储的是对象数组，会导致无法正确查询嵌套字段，需修改为nested类型：

{
  "mappings": {
    "properties": {
      "age": {
        "type": "long"
      },
      "hobbiles": {
        "type": "nested",
        "properties": {
          "name": {
            "type": "keyword"
          },
          "level": {
            "type": "keyword"
          }
        }
      }
    }
  }
}

示例文档 #

[
  {
    "_id": "test@domain.com",
    "age": 12,
    "hobbiles": [
      { "name": "Singing", "level": "begineer" },
      { "name": "Dancing", "level": "begineer" }
    ]
  },
  {
    "_id": "test1@domain.com",
    "age": 7,
    "hobbiles": [
      { "name": "Coding", "level": "begineer" },
      { "name": "Chess", "level": "begineer" }
    ]
  },
  {
    "_id": "test2@domain.com",
    "age": 20,
    "hobbiles": [
      { "name": "Singing", "level": "begineer" },
      { "name": "Dancing", "level": "begineer" }
    ]
  },
  {
    "_id": "test3@domain.com",
    "age": 21,
    "hobbiles": [
      { "name": "Coding", "level": "begineer" },
      { "name": "Dancing", "level": "Football" }
    ]
  }
]

需求说明 #

• 基础条件：文档_id在["test@domain.com", "test1@domain.com", "test3@domain.com"]范围内，且age > 5
• 排序规则：匹配hobbiles.level为Football的文档排在结果顶部，其余满足基础条件的文档紧随其后

查询DSL实现 #

{
  "query": {
    "bool": {
      "filter": [
        { "terms": { "_id": ["test@domain.com", "test1@domain.com", "test3@domain.com"] } },
        { "range": { "age": { "gt": 5 } } }
      ],
      "should": [
        {
          "nested": {
            "path": "hobbiles",
            "query": {
              "term": { "hobbiles.level": "Football" }
            }
          }
        }
      ],
      "minimum_should_match": 0
    }
  },
  "sort": [
    { "_score": { "order": "desc" } },
    { "_id": { "order": "asc" } }
  ]
}

结果说明 #

• 匹配hobbiles.level:Football的test3@domain.com会获得更高评分，排在首位
• 满足基础条件的test@domain.com和test1@domain.com按_id升序排列在后面，最终结果顺序为[test3@domain.com, test@domain.com, test1@domain.com]

内容的提问来源于stack exchange，提问作者Pradip