时间戳

首页 > 解决方案 > ElasticSearch 查询具有给定值或没有值的字段

elasticsearch - ElasticSearch 查询具有给定值或没有值的字段

问题描述

我正在寻找一种方法来查询我的文档以查找符合这些条件的所有文档：

  1. 无论是房产eye.keyword=RIGHT
  2. 或者属性eye.keyword未定义（值缺失/空）

这是我的查询，但我不确定它是否正确并且以最有效的方式编写。我不需要这些结果的分数：

{
  "bool": {
    "should": [
      {
        "terms": {
          "eye.keyword": [
            "RIGHT"
          ],
          "boost": 1
        }
      }
    ],
    "filter": [
      {
        "terms": {
          "indexLocation.keyword": [
            "global",
            "mycompany"
          ],
          "boost": 1
        }
      }
    ],
    "must_not": [
      {
        "exists": {
          "field": "eye.keyword",
          "boost": 1
        }
      }
    ],
    "adjust_pure_negative": true,
    "boost": 1
  }
}

这是一个文档示例：

{
  "_index": "ophthalmiclens",
  "_type": "_doc",
  "_id": "GLOBAL_SP015_0300603412_IT",
  "_version": 3,
  "_score": 1.6931472,
  "_routing": "global",
  "_source": {
    "gradientColor": false,
    "indexLocation": "global",
    "salesPrice": 39.423,
    "description": "MyBrand Addpower 60 1,5 Ø 65 Sph 2,5 Cyl 0 Add 0,75 Dx",
    "range": false,
    "diameterMin": 65,
    "additionMin": 0.75,
    "preset": true,
    "purchasePrice": 8.9,
    "source": "STOCK",
    "type": "DEGRESSIVE",
    "trial": false,
    "manufacturer": "MyBrand",
    "sphereMax": 2.5,
    "lineCode": "ADDPOWER-65",
    "multiCoating": false,
    "cylinderMax": 0,
    "design": "SPHERIC",
    "imageUrl": null,
    "diameterMax": 65,
    "solidColor": false,
    "hardCoating": false,
    "mirroring": false,
    "sku": "0300603412",
    "thumbUrl": null,
    "barcode": "0300603412",
    "prismMax": null,
    "sphereMin": 2.5,
    "coatingCode": "",
    "lineDescription": "Addpower 60",
    "cylinderMin": 0,
    "index": 1.5,
    "photochromic": false,
    "discontinued": false,
    "searchKey": "GLOBAL_SP015_0300603412_IT",
    "prismMin": null,
    "eye": "RIGHT",
    "taxRate": 4,
    "material": "ORGANIC",
    "additionMax": 0.75,
    "polarized": false
  }
}

标签： elasticsearch

解决方案

过滤器块内的任何查询都不会被考虑进行评分。这也称为过滤上下文。所以一切都应该在过滤器块内。查询结构应如下所示

query
  |_ filter
        |_ location = {a, b}
        |_ should
              |_ eye = right
              |_ eye = null (or eye missing)

查询翻译为：

{
  "query": {
    "bool": {
      "filter": [
        {
          "terms": {
            "indexLocation.keyword": [
              "global",
              "mycompany"
            ]
          }
        },
        {
          "bool": {
            "should": [
              {
                "terms": {
                  "eye.keyword": [
                    "RIGHT"
                  ]
                }
              },
              {
                "bool": {
                  "must_not": [
                    {
                      "exists": {
                        "field": "eye"
                      }
                    }
                  ]
                }
              }
            ]
          }
        }
      ]
    }
  }
}

推荐阅读