elasticsearch - 通过减去两个数字字段对elasticsearch进行排序

Question

是否可以通过两个字段之间的减法对结果进行排序？

例如。有两个字段:“价格”，“lastPrice”。我将按价格-lastPrice对文档进行排序。

可以直接完成此操作，还是唯一的方法就是创建一个新字段？

Answer 1

您可以使用脚本排序

数据:

[
      {
        "_index" : "index19",
        "_type" : "_doc",
        "_id" : "LI9zpnEB1jFeu7E5l58r",
        "_score" : 1.0,
        "_source" : {
          "title" : "a",
          "price" : 10,
          "lastPrice" : 15
        }
      },
      {
        "_index" : "index19",
        "_type" : "_doc",
        "_id" : "LY9zpnEB1jFeu7E5xJ_A",
        "_score" : 1.0,
        "_source" : {
          "title" : "b",
          "price" : 15,
          "lastPrice" : 20
        }
      },
      {
        "_index" : "index19",
        "_type" : "_doc",
        "_id" : "Lo95pnEB1jFeu7E5Sp9w",
        "_score" : 1.0,
        "_source" : {
          "title" : "b",
          "price" : 20,
          "lastPrice" : 21
        }
      }
    ]

查询:

{
  "sort": [
    {
      "_script": {
        "type": "number",
        "script": {
          "lang": "painless",
          "source": "doc['lastPrice'].value-doc['price'].value"
        },
        "order": "asc"
      }
    }
  ]
}

结果:

[
      {
        "_index" : "index19",
        "_type" : "_doc",
        "_id" : "Lo95pnEB1jFeu7E5Sp9w",
        "_score" : null,
        "_source" : {
          "title" : "b",
          "price" : 20,
          "lastPrice" : 21
        },
        "sort" : [
          1.0
        ]
      },
      {
        "_index" : "index19",
        "_type" : "_doc",
        "_id" : "LI9zpnEB1jFeu7E5l58r",
        "_score" : null,
        "_source" : {
          "title" : "a",
          "price" : 10,
          "lastPrice" : 15
        },
        "sort" : [
          5.0
        ]
      },
      {
        "_index" : "index19",
        "_type" : "_doc",
        "_id" : "LY9zpnEB1jFeu7E5xJ_A",
        "_score" : null,
        "_source" : {
          "title" : "b",
          "price" : 15,
          "lastPrice" : 20
        },
        "sort" : [
          5.0
        ]
      }
    ]