mongodb 对未填充的子文档进行排序

Question

我正在尝试对未绑定的子文档进行排序。这是合集：

[
    {
        otherFields: {},
        visites: ["58d46ffbf1c383862cf5de4a", "58d46ffbf1c383862cf5de49"]
    }
]

这是合计：

db.exps.aggregate([
    { "$match": {} },
    { "$project": { "visites": 1 } },
    { "$unwind": "$visites" },
    { "$sort": { "visites.formVersion": 1 } },
    { "$group": { "_id": null, "count": { "$sum": 1 }, "data": { "$push": "$visites" } } }
])

我意识到用子字段排序是没有意义的，因为 visites 字段只是一个 ID 数组。是否可以将它们填充到聚合中，以便对它们进行排序？

nodejs填充后的结果是这些：

{
    "count": 50,
    "data": [
        {
            "_id": "58d46ffbf1c383862cf5de4a",
            "formVersion": 4,
            "start": "2017-06-04T07:30:00.000Z",
            "end": "2017-06-04T08:30:00.000Z",
            "type": "assessor",
            "expedient": "58d46ffaf1c383862cf5de41",
            "user": "58d46ff8f1c383862cf5de3a",
            "__v": 0,
            "isDeleted": false,
            "lastModified": "2017-03-24T01:01:47.343Z",
            "created": "2017-03-24T01:01:47.330Z"
        },
        {
            "_id": "58d46ffbf1c383862cf5de49",
            "formVersion": 3,
            "start": "2017-06-03T07:30:00.000Z",
            "end": "2017-06-03T08:30:00.000Z",
            "type": "assessor",
            "expedient": "58d46ffaf1c383862cf5de41",
            "user": "58d46ff8f1c383862cf5de3a",
            "__v": 0,
            "isDeleted": false,
            "lastModified": "2017-03-24T01:01:47.341Z",
            "created": "2017-03-24T01:01:47.328Z"
        },

Answer 1

假设两个集合 coll1 和 coll2:

> db.coll1.find()

{
 "_id": ObjectId("58eecf2a11b7af6f560666d8"),
  "otherFields": {

 },
 "visites": [
   "58d46ffbf1c383862cf5de4a",
   "58d46ffbf1c383862cf5de49"
  ]
}
> db.coll2.find()
{
  "_id": "58d46ffbf1c383862cf5de4a",
  "formVersion": 4,
  "start": "2017-06-04T07:30:00.000Z",
  "end": "2017-06-04T08:30:00.000Z",
  "type": "assessor",
  "expedient": "58d46ffaf1c383862cf5de41",
  "user": "58d46ff8f1c383862cf5de3a",
  "__v": 0,
  "isDeleted": false,
  "lastModified": "2017-03-24T01:01:47.343Z",
  "created": "2017-03-24T01:01:47.330Z"
}
{
  "_id": "58d46ffbf1c383862cf5de49",
  "formVersion": 3,
  "start": "2017-06-03T07:30:00.000Z",
  "end": "2017-06-03T08:30:00.000Z",
  "type": "assessor",
  "expedient": "58d46ffaf1c383862cf5de41",
  "user": "58d46ff8f1c383862cf5de3a",
  "__v": 0,
  "isDeleted": false,
  "lastModified": "2017-03-24T01:01:47.341Z",
  "created": "2017-03-24T01:01:47.328Z"
}

如果您希望按例如formVersion 字段，一种可能的聚合管道解决方案可能是： db.coll1.aggregate([

// left outer join from coll2 to coll1
{$lookup:{
    from:"coll2",
    localField:"visites",
    foreignField:"_id",
    as:"data"
}},

// unwind the resulting array
{$unwind:"$data"},

// sort the unwinded documents using the formVersion field
{$sort:{"_id":1 ,"data.formVersion": 1}},

// collect the results into an array
{$group:{"_id":null, data:{$push:"$data"}}}
])

流水线的结果是：

 {
  "_id": null,
  "data": [
    {
      "_id": "58d46ffbf1c383862cf5de49",
      "formVersion": 3,
      "start": "2017-06-03T07:30:00.000Z",
      "end": "2017-06-03T08:30:00.000Z",
      "type": "assessor",
      "expedient": "58d46ffaf1c383862cf5de41",
      "user": "58d46ff8f1c383862cf5de3a",
      "__v": 0,
      "isDeleted": false,
      "lastModified": "2017-03-24T01:01:47.341Z",
      "created": "2017-03-24T01:01:47.328Z"
    },
    {
      "_id": "58d46ffbf1c383862cf5de4a",
      "formVersion": 4,
      "start": "2017-06-04T07:30:00.000Z",
      "end": "2017-06-04T08:30:00.000Z",
      "type": "assessor",
      "expedient": "58d46ffaf1c383862cf5de41",
      "user": "58d46ff8f1c383862cf5de3a",
      "__v": 0,
      "isDeleted": false,
      "lastModified": "2017-03-24T01:01:47.343Z",
      "created": "2017-03-24T01:01:47.330Z"
    }
  ]
}

请注意，从 MongoDB 3.4 开始，$lookup 阶段存在某些限制；比如两个集合必须在同一个数据库中，并且from参数不能是分片集合。详情请见https://docs.mongodb.com/manual/reference/operator/aggregation/lookup/ 另请注意，最后的 $group 阶段将生成单个文档，该文档将受到 MongoDB 的 16 MB 文档大小限制的约束。如果您愿意，可以删除 $group 阶段并迭代 $sort 阶段返回的游标。

mongodb 对未填充的子文档进行排序

mongodb sort unpopulated subdocument

sorting

populate

mongodb

aggregation-framework