如何确保无痛脚本不会导致重新编译？

Question

我在排序上下文中使用脚本来实现自定义排序。以下是我的脚本排序节点，允许自定义排序。

{
  "_script": {
    "type": "string",
    "script": {
      "source": """
            
                   def sortingValue = "";
        for(query in params.queries) {
              if (doc['firstName.keyword'].size() > 0 && doc['firstName.keyword'].value.startsWith(query)) {
                   sortingValue = '1' +":"+ doc['fullName.keyword'].value;
                   break;
               } else if (doc['lastName.keyword'].size() > 0 && doc['lastName.keyword'].value.startsWith(query)) {
                   sortingValue = '2' +":"+ doc['fullName.keyword'].value;
                   break;
                } else if (doc['companyName.keyword'].size() > 0 && doc['companyName.keyword'].value.contains(query)) {
                   sortingValue = '3' + ":" + doc['companyName.keyword'].value;
                   break;
                } else {
                   sortingValue = '4';
                }
        }
        return sortingValue;
                """,
      "params": {
        "queries": [
          "mo",
          "buga"
        ]
      }
    },
    "order": "asc"
  }
}

使用脚本，我按以下方式对结果进行分类

搜索文字：莫布加

第一类： 对名字以 Mo 或 Buga 开头的联系人进行排序

第二类： 对姓氏以 Mo 或 Buga 开头的联系人进行排序

第三类： 将公司名称以 Mo 或 Buga 开头的联系人排序为第三类

第四类： 所有其他匹配结果都属于第四类

我如何确保没有不需要的重新编译。

Answer 1

您可以通过

edge_ngram

过滤器和

multi_match

查询对没有脚本的点击进行排序

映射

PUT /first_last_company_name
{
    "mappings": {
        "properties": {
            "first_name": {
                "type": "text",
                "analyzer": "whitespace_edge_ngram_filter_analyzer"
            },
            "last_name": {
                "type": "text",
                "analyzer": "whitespace_edge_ngram_filter_analyzer"
            },
            "company_name": {
                "type": "text",
                "analyzer": "whitespace_edge_ngram_filter_analyzer"
            }
        }
    },
    "settings": {
        "analysis": {
            "analyzer": {
                "whitespace_edge_ngram_filter_analyzer": {
                    "tokenizer": "whitespace",
                    "filter": [
                        "lowercase",
                        "edge_ngram_filter"
                    ]
                }
            },
            "filter": {
                "edge_ngram_filter": {
                    "type": "edge_ngram",
                    "min_gram": 1,
                    "max_gram": 11
                }
            }
        }
    }
}

具有不同名字、姓氏和公司名称组合的文件

PUT /first_last_company_name/_bulk
{"create":{"_id":1}}
{"first_name":"mogun","last_name":"nomo","company_name":"nobuga"}
{"create":{"_id":2}}
{"first_name":"bugatti","last_name":"nomo","company_name":"nobuga"}
{"create":{"_id":3}}
{"first_name":"nomon","last_name":"motto","company_name":"nobuga"}
{"create":{"_id":4}}
{"first_name":"nomon","last_name":"bugar","company_name":"nobuga"}
{"create":{"_id":5}}
{"first_name":"nomon","last_name":"nobugam","company_name":"more"}
{"create":{"_id":6}}
{"first_name":"nomon","last_name":"nobugam","company_name":"bugastein"}
{"create":{"_id":7}}
{"first_name":"nomon","last_name":"bugav","company_name":"monoq"}
{"create":{"_id":8}}
{"first_name":"nomon","last_name":"mower","company_name":"bugano"}
{"create":{"_id":9}}
{"first_name":"nomon","last_name":"nobuga","company_name":"nomobuga"}

multi_match

带响应过滤器的查询

GET /first_last_company_name/_search?filter_path=hits.hits._source, hits.hits._score,hits.hits._id
{
    "query": {
        "multi_match": {
            "query": "mo buga",
            "tie_breaker": 0,
            "fields": [
                "first_name^100",
                "last_name^10",
                "company_name"
            ]
        }
    }
}

回应

{
    "hits" : {
        "hits" : [
            {
                "_id" : "2",
                "_score" : 373.28677,
                "_source" : {
                    "first_name" : "bugatti",
                    "last_name" : "nomo",
                    "company_name" : "nobuga"
                }
            },
            {
                "_id" : "1",
                "_score" : 337.62857,
                "_source" : {
                    "first_name" : "mogun",
                    "last_name" : "nomo",
                    "company_name" : "nobuga"
                }
            },
            {
                "_id" : "4",
                "_score" : 27.299335,
                "_source" : {
                    "first_name" : "nomon",
                    "last_name" : "bugar",
                    "company_name" : "nobuga"
                }
            },
            {
                "_id" : "7",
                "_score" : 27.299335,
                "_source" : {
                    "first_name" : "nomon",
                    "last_name" : "bugav",
                    "company_name" : "monoq"
                }
            },
            {
                "_id" : "3",
                "_score" : 24.707626,
                "_source" : {
                    "first_name" : "nomon",
                    "last_name" : "motto",
                    "company_name" : "nobuga"
                }
            },
            {
                "_id" : "8",
                "_score" : 24.707626,
                "_source" : {
                    "first_name" : "nomon",
                    "last_name" : "mower",
                    "company_name" : "bugano"
                }
            },
            {
                "_id" : "6",
                "_score" : 2.7447402,
                "_source" : {
                    "first_name" : "nomon",
                    "last_name" : "nobugam",
                    "company_name" : "bugastein"
                }
            },
            {
                "_id" : "5",
                "_score" : 2.4951272,
                "_source" : {
                    "first_name" : "nomon",
                    "last_name" : "nobugam",
                    "company_name" : "more"
                }
            }
        ]
    }
}

_score

在匹配名字、姓氏或公司名称方面存在显着差异