如何提高结果在elasticsearch，使匹配在field1总是高于匹配在field2

Question

有映射，3个字段和9个文档:

#! /bin/bash
#DELETE
curl -XDELETE 'http://localhost:9200/test'
echo
# CREATE
curl -XPUT 'http://localhost:9200/test?pretty=1' -d '{
    "settings": {
       "analysis" : {
            "analyzer" : {
               "my_analyz_1" : {
                    "filter" : [
                        "standard",
                        "lowercase",
                        "asciifolding"
                    ],
                    "type" : "custom",
                    "tokenizer" : "standard"
                }
            }
        }
    }
}'
echo
# DEFINE
curl -XPUT 'http://localhost:9200/test/posts/_mapping?pretty=1' -d '{
    "posts" : {
        "properties" : {
            "section" : {
              "type" : "string",
              "analyzer" : "my_analyz_1"
            },
            "category" : {
              "type" : "string",
              "analyzer" : "my_analyz_1"
            },
            "title" : {
              "type" : "string",
              "analyzer" : "my_analyz_1"
            }
        }
    }
}'
echo
# INSERT
curl localhost:9200/test/posts/1 -d '{section: "Bicycle", category: "Small", title: "Diamondback Grind-16"}'
curl localhost:9200/test/posts/2 -d '{section: "Bicycle", category: "Big",   title: "Diamondback JrViper"}'
curl localhost:9200/test/posts/3 -d '{section: "Bicycle", category: "Small", title: "2-Hip Cyclone small"}'
curl localhost:9200/test/posts/4 -d '{section: "Bicycle", category: "Big",   title: "2-Hip Bizzle"}'
curl localhost:9200/test/posts/5 -d '{section: "Small",   category: "Small", title: "Toyota"}'
curl localhost:9200/test/posts/6 -d '{section: "Car",     category: "Big",   title: "Subaru Impreza small"}'
curl localhost:9200/test/posts/7 -d '{section: "Small",   category: "Big",   title: "Toyota Corona MARK II"}'
curl localhost:9200/test/posts/8 -d '{section: "Car",     category: "Small", title: "Hyundai Elantra"}'
curl localhost:9200/test/posts/9 -d '{section: "Car",     category: "Big",   title: "Ford Maverick small"}'
echo
# REFRESH
curl -XPOST localhost:9200/test/_refresh
echo

我想搜索单词"small"，但是我总是希望结果的顺序如下:

1. 小节
中出现"small"的结果
2. 分类
中"small"的结果
3. 标题
中出现'small'的结果

所以我搜索query:

curl "localhost:9200/test/posts/_search?pretty=1" -d '{
    "query": {
        "bool": {
            "must": [
                {
                    "multi_match": {
                        "query": "small",
                        "fields": ["section^3", "category^2", "title"]
                    }
                }
            ]
        }
    }
}'

，结果为:

{"_id": 7} {section: "Small",   category: "Big",   title: "Toyota Corona MARK II"}
{"_id": 1} {section: "Bicycle", category: "Small", title: "Diamondback Grind-16"}
{"_id": 5} {section: "Small",   category: "Small", title: "Toyota"}
{"_id": 3} {section: "Bicycle", category: "Small", title: "2-Hip Cyclone small"}
{"_id": 8} {section: "Car",     category: "Small", title: "Hyundai Elantra"}
{"_id": 9} {section: "Car",     category: "Big",   title: "Ford Maverick small"
{"_id": 6} {section: "Car",     category: "Big",   title: "Subaru Impreza small"}

这不是我想要的。5应该是第二个，因为match在section中。3应该在7和5之后，因为match在category和title中。

所以，我的问题是，我如何才能获得匹配在section 总是更重要的结果，然后匹配在类别，这是总是比匹配在标题。

提前感谢!

编辑:

使用搜索类型'dfs_query_then_fetch'解决了问题，它计算所有分片的TF-IDF值。更多信息请参见http://www.elasticsearch.org/guide/reference/api/search/search-type/

Answer 1

您是否尝试将use_dis_max设置为false ?

这应该意味着在category和title中带有"small"的文档将高于仅在category中带有"small"的文档。

至于你在第二和第三个结果之间看到的奇怪行为，我有点迷路了…你能做这个查询并问一下吗解释分数是如何计算的?