MongoDB聚合 - 将项目字段值作为字段

Question

在将数据变成棘手的格式时遇到一些麻烦.

文件包含以下字段: subject, importance [high,medium,low]

样本文件:

{_id: "", subject: "red", importance: "high"}

我想返回看起来像这样的数据:

[{_id: subject, high: 5, medium: 6, low: 3}] 

这些数字与每个重要性级别的文档数量相对应

这是我到目前为止:

{
$group: {
  _id: {subject: "$subject", importance: "$importance"},
  count: {$sum: 1},
  }
 },
 {
   $group: {
     _id: "$_id.subject",
     data: {
       $push: {
          importance: "$_id.importance", count: "$count"
       }
     }
   }
 }

有没有办法可以使用值$_id.importance并将其作为键并具有$count值？

Answer 1

您可以按如下方式汇总：

• $group通过subject和importance，得到各自的计数。
• 然后是棘手的部分，即条件$project，它将随着字段可以容纳的选项数量线性增长importance。目前是三 -high和low。medium
• $group结果再次返回subject，并使用$sum运算符accumulate对重要性字段的不同值进行计数。

示例代码：

db.t.aggregate([
{$group:{"_id":{"subject":"$subject",
                "importance":"$importance"},
         "count":{$sum:1}}},
{$project:{"_id":0,
           "subject":"$_id.subject",
           "result":{$cond:[
                           {$eq:["$_id.importance","high"]},
                           {"high":"$count"},
                           {$cond:[{$eq:["$_id.importance","low"]},
                                   {"low":"$count"},
                                   {"medium":"$count"}]}]}}},
{$group:{"_id":"$subject",
         "low":{$sum:"$result.low"},
         "medium":{$sum:"$result.medium"},
         "high":{$sum:"$result.high"}}},
])

测试数据：

db.t.insert([
{"subject":"history","importance":"high"},
{"subject":"geography","importance":"low"},
{"subject":"history","importance":"low"},
{"subject":"history","importance":"medium"},
{"subject":"geography","importance":"low"},
{"subject":"history","importance":"low"}
])

结果：

{ "_id" : "geography", "low" : 2, "medium" : 0, "high" : 0 }
{ "_id" : "history", "low" : 2, "medium" : 1, "high" : 1 }