ElasticSearch -2域上的复杂序？

Question

我有这个样本数据

Name: A
Price: 200
PromotionPrice: 100
PromotionStart: 2020/01/01
PromotionEnd:   2020/01/15

如果没有晋升，PromotionPrice将为0。

我想把那些有PromotionPrice先开始的(asc或desc)和那些没有PromotionPrice的排序。(过滤掉那些有PromotionPrice但超出促销时间范围)

我试着在_script上工作，但这对我来说太复杂了。

目前正在开发版本7。

我的地图：

"Price" : {"type": "long"}
"PromotionPrice": {"type": "long"},
"PromotionStart": {"type": "date"},  
"PromotionEnd": {"type": "date"},

我的代码:按Price或PromotionPrice排序，但没有优先考虑那些项目和升级。

"sort": [
            {
                "_script": {
                    "type": "number",
                    "script": {
                        "lang": "painless",
                        "params": {
                            "now": "120000000"
                        },
                        "source": "if (doc['PromotionPrice'].value != 0 && params.now <= doc['PromotionEnd'].value.getMillis() && params.now >= doc['PromotionStart'].value.getMillis()) {doc['selling_price'].value} else {doc['normal_price'].value}"
                    },
                    "order": "asc"
                }
            }
        ]

预期结果：

Name: A
Price: 200
PromotionPrice: 100
PromotionStart: 2020/11/01
PromotionEnd:   2020/12/01

Name: B
Price: 500
PromotionPrice: 200
PromotionStart: 2020/11/01
PromotionEnd:   2020/12/01


Name: C
Price: 500
PromotionPrice: 300
PromotionStart: 2020/11/01
PromotionEnd:   2020/12/01


Name: E
Price: 100
PromotionPrice: 0
PromotionStart: 2020/11/01
PromotionEnd:   2020/12/01


Name: D                              
Price: 500               #have PromotionPrice but treated as not having 
PromotionPrice: 300      #cuz out of promotion time range
PromotionStart: 2020/11/01
PromotionEnd:   2020/11/05

进一步专题：

基于@Val的逻辑：

如果我想对asc或desc进行排序，下面是那些有晋升的：

now和type将包含在params中

def isPromo = (doc.new_price.value > 0 && doc.start_date.value.getMillis() <  Long.parseLong(params.now) && doc.end_date.value.getMillis() >  Long.parseLong(params.now)); 
if (isPromo? && if (params.type == 'asc') {return true;} return false;) 
(return doc.new_price.value: (doc.price.value+100000000) return doc.new_price.value: (doc.price.value-100000000

遗憾的是这个不起作用

Answer 1

很好的开始，您的脚本几乎包含正确的逻辑，您只需要添加一个常数到常规价格，您将在下面的反应中看到。

下面的脚本可以帮助您按照预期的方式对结果进行排序。它的工作方式包含在脚本中的注释中：

POST promotions/_search
{
  "sort": {
    "_script": {
      "type": "number",
      "script": {
        "lang": "painless",
        "source": """
        // current time
        def now = new Date().getTime();

        // check if the current product satisfies to the promotion constraints
        def isPromo = (doc.PromotionPrice.value > 0 && doc.PromotionStart.value.getMillis() < now && doc.PromotionEnd.value.getMillis() > now);

        // if the product qualifies for promotion return its promotion price, otherwise return its normal price + a constant
        return isPromo ? doc.PromotionPrice.value : (doc.Price.value + 100000000);
        """
      },
      "order": "asc"
    }
  }
}

结果：

    "hits" : [
      {
        "_index" : "promotions",
        "_type" : "_doc",
        "_id" : "ilucrHUB7CBHpnCGfnBn",
        "_score" : null,
        "_source" : {
          "Name" : "A",
          "Price" : 200,
          "PromotionPrice" : 100,
          "PromotionStart" : "2020-11-01",
          "PromotionEnd" : "2020-12-01"
        },
        "sort" : [
          100.0
        ]
      },
      {
        "_index" : "promotions",
        "_type" : "_doc",
        "_id" : "i1ucrHUB7CBHpnCGfnBn",
        "_score" : null,
        "_source" : {
          "Name" : "B",
          "Price" : 500,
          "PromotionPrice" : 200,
          "PromotionStart" : "2020-11-01",
          "PromotionEnd" : "2020-12-01"
        },
        "sort" : [
          200.0
        ]
      },
      {
        "_index" : "promotions",
        "_type" : "_doc",
        "_id" : "jFucrHUB7CBHpnCGfnBn",
        "_score" : null,
        "_source" : {
          "Name" : "C",
          "Price" : 500,
          "PromotionPrice" : 300,
          "PromotionStart" : "2020-11-01",
          "PromotionEnd" : "2020-12-01"
        },
        "sort" : [
          300.0
        ]
      },
      {
        "_index" : "promotions",
        "_type" : "_doc",
        "_id" : "jVucrHUB7CBHpnCGfnBn",
        "_score" : null,
        "_source" : {
          "Name" : "E",
          "Price" : 100,
          "PromotionPrice" : 0,
          "PromotionStart" : "2020-11-01",
          "PromotionEnd" : "2020-12-01"
        },
        "sort" : [
          1.000001E8
        ]
      },
      {
        "_index" : "promotions",
        "_type" : "_doc",
        "_id" : "jlucrHUB7CBHpnCGfnBn",
        "_score" : null,
        "_source" : {
          "Name" : "D",
          "Price" : 500,
          "PromotionPrice" : 300,
          "PromotionStart" : "2020-11-01",
          "PromotionEnd" : "2020-11-05"
        },
        "sort" : [
          1.000005E8
        ]
      }
    ]