在 multi_match 查询中，是否有可能以与字段相关性类似的方式提高术语的相关性？

Question

所以我们知道可以通过使用插入符号 (^) 后跟数字来提高字段的相关性，如下所示：

{
    "query": {
        "multi_match": {
            "query": "something of value",
            "fields": [ "name^2", "category" ]
        }
    }
}

但对术语 (query) 应用相同的逻辑并不能提高其相关性：

{
    "query": {
        "multi_match": {
            "query": "something^3 of value",
            "fields": [ "name", "category" ]
        }
    }
}

我想要做的是提升query 的第一个单词，因此希望在fields 的每个字段开头包含该单词的文档优先。

我用span_term (https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-span-term-query.html) 尝试了一些东西，但它不支持multi_match。

有什么想法吗？

Answer 1

简而言之，回答您的问题：不，使用多重匹配查询是不可能的。一种提升词条的方法是查询字符串查询。

TLDR；

我认为您对多重匹配查询有误解。

您可以使用插入符号来提升字段，而不是查询本身。另见issue。

您可以改用Query String Query

我创建了两个文档

PUT test/_doc/5
{
  "name": "Something Blue Something Borrowed Something New",
  "category": "Comedy"
}
PUT test/_doc/6
{
  "name": "There's Something About Mary",
  "category": "Comedy"
}

搜索，提升词条蓝色

GET test/_search
{
    "query": {
        "query_string" : {
            "default_field" : "name",
            "query" : "Something Blue^2"
        }
    }
}

搜索结果

{
  "took" : 4,
  "hits" : {
    "total" : 2,
    "max_score" : 1.027436,
    "hits" : [
      {
        "_index" : "test",
        "_type" : "_doc",
        "_id" : "5",
        "_score" : 1.027436,
        "_source" : {
          "name" : "Something Blue Something Borrowed Something New",
          "category" : "Comedy"
        }
      },
      {
        "_index" : "test",
        "_type" : "_doc",
        "_id" : "6",
        "_score" : 0.7590336,
        "_source" : {
          "name" : "There's Something About Mary",
          "category" : "Comedy"
        }
      }
    ]
  }
}

如果你问解释 api

GET test/_doc/5/_explain
{
    "query": {
        "query_string" : {
            "default_field" : "name",
            "query" : "Something Blue^2"
        }
    }
}

相关的解释结果

{
        "value" : 0.5753642,
        "description" : "weight(name:blue in 0) [PerFieldSimilarity], result of:",
        "details" : [
          {
            "value" : 0.5753642,
            "description" : "score(doc=0,freq=1.0 = termFreq=1.0\n), product of:",
            "details" : [
              {
                "value" : 2.0,
                "description" : "boost",
                "details" : [ ]
              },
...