Mongodb聚合以对数组中最常见的项目进行排序？答案

【问题标题】：Mongodb aggregate to sort items with most common items in array?Mongodb聚合以对数组中最常见的项目进行排序？
【发布时间】：2020-11-02 17:36:12
【问题描述】：

集合中的一个项目包含一个字符串数组。我想查找数组中匹配元素最多的项目并对其进行排序。

考虑一个集合：

[
    {
        "item_name":"Item_1",
        "tags":["A","B","C","D","E"]
    },
    {
        "item_name":"Item_2",
        "tags":["A","B","D","E","G"]
    },
    {
        "item_name":"Item_3",
        "tags":["B","C","E","H"]
    }

]

我想根据 ["B","D","G","F"] 之类的数组对集合进行排序这将返回

[

    {
        "item_name":"Item_2",
        "tags":["A","B","D","E","G"]
    },
    {
        "item_name":"Item_1",
        "tags":["A","B","C","D","E"]
    },
    {
        "item_name":"Item_3",
        "tags":["B","C","E","H"]
    }

]

预期的顺序是 Item_2,Item_1，然后是 Item_3，

Item_2 匹配 3 个项目（“B”、“D”和“G”）
然后 Item_1 有 2 个匹配项（“B”和“D”）
最后是 Item_3 与 1 个匹配项 ("B")

如果不在 mongodb 中，JavaScript 方法也会受到赞赏

【问题讨论】：

标签： javascript mongodb mongoose mongodb-query aggregation-framework

【解决方案1】：

您可以使用$setIntersection 获取相交项目的数量，然后根据计算的分数使用$sort。

db.collection.aggregate([
  {
    $project: {
      _id: 0,
      item_name: "$item_name",
      tags: "$tags",
      score: {
        $let: {
          vars: {
            intersection: {
              $setIntersection: [
                "$tags",
                [
                  "B",
                  "D",
                  "G",
                  "F"
                ]
              ]
            }
          },
          in: {
            $size: "$$intersection"
          }
        }
      }
    }
  },
  {
    $sort: {
      score: -1
    }
  }
])

https://mongoplayground.net/p/tpDTtVKetFT

【讨论】：

【解决方案2】：

如果tags 中的值是唯一的，则可以使用tags 的交集大小和使用$size 和$setIntersection 的查询数组

db.collection.aggregate([
  {
    $set: {
      matchedCount: {
        $size: {
          $setIntersection: ["$tags", ["B","D","G","F"]]
        }
      }
    }
  },
  {
    $sort: {
      matchedCount: -1
    }
  }
])

【讨论】：

谢谢。这是准确的。您能否建议如何丢弃与至少 1 个标签不匹配的项目
假设您正在搜索 ["B","D","G","F"]，当您说“不匹配至少 1 个标签的项目”时，这是否意味着 ["A","D","G","F"] 或 ["B","D","G","F","H"] 或 ["D"," G","F"]?
我的意思是 ["H","K","L"]
您可以在管道的开头添加{ $match: { tags: { $in: ["B","D","G","F"] } } }