multi_match 的 minimum_should_match

minimum_should_match 只能用在multi_match

可以这样写:

{
  "query": {
        "multi_match": {
          "query":"徐峥沈腾黄渤吃火锅",
          "fields":["title"],
          "minimum_should_match":4
        }
    }
}

表示命中4个词的文档才会返回

也可以这样写:

{
  "query": {
        "multi_match": {
          "query":"徐峥沈腾黄渤吃火锅",
          "fields":["title"],
          "minimum_should_match":"80%"
        }
    }
}

这里的80%是指query的分词个数的80%

我们调一下分词接口,看结果:

{
    "tokens": [
        {
            "token": "徐峥",
            "start_offset": 0,
            "end_offset": 2,
            "type": "CN_WORD",
            "position": 0
        },
        {
            "token": "沈腾",
            "start_offset": 2,
            "end_offset": 4,
            "type": "CN_WORD",
            "position": 1
        },
        {
            "token": "黄渤",
            "start_offset": 4,
            "end_offset": 6,
            "type": "CN_WORD",
            "position": 2
        },
        {
            "token": "吃火锅",
            "start_offset": 6,
            "end_offset": 9,
            "type": "CN_WORD",
            "position": 3
        }
    ]
}

query被分成了4个词,所以minimum_should_match:"80%" 在这里等于 minimum_should_match:4*0.8

还要注意下,这里的小数值会向下取整,也就是命中3个词的就会返回。

原文地址:https://www.cnblogs.com/feng07/p/11571152.html