在MongoDB的聚合管道中获取$group之后输入文档中的字段

Question

我面临的问题是如何在一次组操作后访问原始文档，并在MongoDB的聚合流水线中携带$group后的字段。

原始文件是：

{
"_id" : ObjectId("361de42f1938e89b179dda42"),
"user_id" : ObjectId("9424021bafbde55512e39b83"),
"candidate_id" : ObjectId("54f65356294160421ead3ca1")
"OVERALL_SCORE" : 150,
"SCORES" : [ 
    { "NAME" : "asd", "OBTAINED_SCORE" : 28}, { "NAME" : "acd", "OBTAINED_SCORE" : 36 }, { "NAME" : "abc", "OBTAINED_SCORE" : 40}
 ]
}

聚合功能：

 db.coll.aggregate([ $group : { _id : { user_id : "$user_id"}, BEST_SCORE : { $max : "$OVERALL_SCORE"}, AVG_SCORE : { $avg : "$OVERALL_SCORE" }}} ])

下面是示例输出（第1组之后）：

{
"result" : [ 
    {
        "_id" : {
            "user_id" : ObjectId("9424021bafbde55512e39b83")
        },
        "BEST_SCORE" : 150,
        "AVG_SCORE" : 132
    }
],
"ok" : 1
 }

   db.coll.aggregate([ $group : { _id : { user_id : "$user_id"}, BEST_SCORE : { $max : "$OVERALL_SCORE"}, AVG_SCORE : { $avg : "$OVERALL_SCORE" }}}, { $unwind : "$SCORES"}, /*problem is--after group operation "SCORES" field which is in original document not available */ { $group : _id : { NAME: "$SCORES.NAME"}, AVG_OBTAINED_SCORE: { $avg : "$SCORES.OBTAINED_SCORE"}} **/*problem is--this is also in the original document*/** ])

   "BEST_SCORE": 150,                     //after 1st group
  "AVG_SCORE": 132,                       //after 1st group
  "SCORES": [                             //problem --- unwind "SCORES" and then group which is actually will not be available after 1st group (get this from original document)
    {
      "NAME": "abc",
      "AVG_OBTAINED_SCORE": 25.5
    },
    {
      "NAME": "asd",
      "AVG_OBTAINED_SCORE": 24
    },
    {
      "NAME": "acd",
      "AVG_OBTAINED_SCORE": 32
    }
  ]

谢谢。

Answer 1

当使用要保留组中所有考虑的文档的值的内容进行分组时，需要使用$push。关键在于，这是一个数组。因此，您处理$unwind两次，并且还有两个$group阶段：

db.coll.aggregate([
    {  "$group" : { 
        "_id": "$user_id", 
        "BEST_SCORE": { "$max": "$OVERALL_SCORE" },
        "AVG_SCORE": { "$avg": "$OVERALL_SCORE" },
        "SCORES": { "$push": "SCORES" }
    }}, 

    // SCORES in an array of arrays. Unwind twice
    { "$unwind": "$SCORES" },
    { "$unwind": "$SCORES" },

    // Group for averages on elements
    { "$group": {
        "_id": {
            "user_id": "$_id",
            "NAME": "$SCORES.name"
        },
        "BEST_SCORE": { "$first": "$BEST_SCORE" },
        "AVG_SCORE": { "$first": "$AVG_SCORE" }
        "AVG_OBTAINED_SCORE": { "$avg": "$SCORES.OBTAINED_SCORE" } 
    }},

    // Group to user_id
    { "$group": {
        "user_id": "$_id.user_id",
        "BEST_SCORE": { "$first": "$BEST_SCORE" },
        "AVG_SCORE": { "$first": "$AVG_SCORE" }
        "SCORES": { "$push": {
            "NAME": "$_id.NAME",
            "AVG_OBTAINED_SCORE": "$AVG_OBTAINED_SCORE"
        }}     
    }}
])

您可能会考虑在第一个$group之前使用$unwind，但是如果这样做了，那么计算的平均值将受到数组中被“取消缠绕”的元素数量的影响。因此，“双美元释放”是一个必要的过程。