管理精确搜索和普通搜索索引的最佳方法是什么

发布于 2025-01-10 20:05:40 字数 752 浏览 2 评论 0原文

我有一个用例，我需要正常的 match_all 搜索，即它也向我显示具有模糊结果的相关文档，但是我需要一个需要执行术语查询的特定搜索。我的映射主要是字符串类型：

"project_title": {
        "type": "text",
        "fields": {
          "keyword": {
            "type": "keyword",
            "ignore_above": 256
          }
        }
      },
      "projectabstract": {
        "type": "text",
        "fields": {
          "keyword": {
            "type": "keyword",
            "ignore_above": 256
          }
        }
      },
"task_number": {
        "type": "keyword",
        "normalizer": "lowercase_normalizer"
      }

我确实通过将 task_number 设置为关键字来尝试使用小写规范化器进行精确搜索，但发现我无法进行模糊搜索，因为它是一个关键字。由于小写规范化器仅适用于关键字，因此如果我将字段转换为关键字，我将无法获得从 type: string 获得的搜索结果。我应该维护两个索引，一个包含字符串，另一个包含关键字类型，还是有更好的替代方案？

原文

I have a use case where I need normal match_all search i.e it shows me related documents with fuzzy results also, however I need a specific search which needs to do a term query.
My mappings are mostly string type:

"project_title": {
        "type": "text",
        "fields": {
          "keyword": {
            "type": "keyword",
            "ignore_above": 256
          }
        }
      },
      "projectabstract": {
        "type": "text",
        "fields": {
          "keyword": {
            "type": "keyword",
            "ignore_above": 256
          }
        }
      },
"task_number": {
        "type": "keyword",
        "normalizer": "lowercase_normalizer"
      }

I did try exact search using lowercase normalizer by making task_number a keyword, but found out that I cant do a fuzzy search because it is a keyword. Since lowercase normalizer works only on keywords, if I convert the fields to keyword I cannot get search results I get from type: string. Should I maintain two indexes one with strings and another of type keyword or is there a better alternative to it ?

分享到QQ

分享到微博