Question

<$>在$ lookup之后我会得到一个名为lastViewed的连接：

{
  "_id" : "5955ea4fd8099718330ab191"

    lastViewed: [
        {
           "_id" : ObjectId("595218a7d346d27fb0bc1705"),
           "userId" : ObjectId("58c796d4344b9da4dfbe027b"),
           "groupId" : ObjectId("5955ea4fd8099718330ab191"),
           "lastViewedTime" : ISODate("2017-06-19T09:39:07.374Z")
        },
        {
           "_id" : ObjectId("595218a7d346d27fb0bc1764"),
           "userId" : ObjectId("58c796d4344b9da4dfbe027b"),
           "groupId" : ObjectId("5955ea4fd8099718330ab162"),
          "lastViewedTime" : ISODate("2017-05-11T09:39:07.374Z")
        }
      ]
}

我想使用$ filter，然后使用$ project值，我的查询位于userId从params获取的位置

$project: { "viewedDetails": {
                        "$filter": {
                            "input": "$lastViewed",
                            "as": "lastViewed",
                            "cond": { $and: [{ "$eq": ["$$lastViewed.userId", mongoose.Types.ObjectId(userId)] }, { "$eq": ["$$lastViewed.groupId", '$_id'] }] }
                        }
                    }
                }

对于userId 58c796d4344b9da4dfbe027b，我得到了如下的确切输出

{
  "_id" : "5955ea4fd8099718330ab191"

    viewedDetails: [
        {
           "_id" : ObjectId("595218a7d346d27fb0bc1705"),
           "userId" : ObjectId("58c796d4344b9da4dfbe027b"),
           "groupId" : ObjectId("5955ea4fd8099718330ab191"),
           "lastViewedTime" : ISODate("2017-06-19T09:39:07.374Z")
        }
      ]
}

现在我将我的架构更改为并将$ lookup结果作为

{
  "_id" : "5955ea4fd8099718330ab191"

    lastViewed: [
        {
           "_id" : ObjectId("595218a7d346d27fb0bc1705"),
           "userId" : ObjectId("58c796d4344b9da4dfbe027b"),
           "members":[
                  {
                    "groupId" : ObjectId("5955ea4fd8099718330ab162"),
                    "lastViewedTime" : ISODate("2017-05-11T09:39:07.374Z")
                  }
                  {
                    "groupId" : ObjectId("5955ea4fd8099718330ab191"),
                    "lastViewedTime" : ISODate("2016-05-19T09:39:07.374Z")
                  }
              ]               
        }
      ]
}

然后我的对象数组的过滤器不起作用，我的查询是

 $project: { "viewedDetails": {
                        "$filter": {
                            "input": "$lastViewed",
                            "as": "lastViewed",
                            "cond": { $and: [{ "$eq": ["$$lastViewed.userId", mongoose.Types.ObjectId(userId)] }, 
                                             { "$eq": ["$$lastViewed.members.groupId", '$_id'] }] }
                        }
                    }
                },

如何在不使用$ unwind的情况下获得此功能，然后使用$ project。

Answer 1

不需要$unwind纯粹用于比较，因为有各种运算符可以通过与列表进行比较来确定逻辑结果。您真的只需更改此处$filter上的cond;

过滤“成员”需要使用$map换行以反映更改：

对于MongoDB 3.4，您可以使用$in：

"viewedDetails": {
  "$map": {
    "input": {
      "$filter": {
        "input": "$lastViewed",
        "as": "lastViewed",
        "cond": { 
          "$and": [
            { "$eq": ["$$lastViewed.userId", mongoose.Types.ObjectId(userId)] }, 
            { "$in": [ "$_id", "$$lastViewed.members.groupId" ] }
          ]
        }
      }
    },
    "as": "v",
    "in": {
      "_id": "$$v._id",
      "userId": "$$v.userId",
      "members": { 
        "$filter": {
          "input": "$$v.members",
          "as": "m",
          "cond": { "$eq": [ "$$m.groupId", "$_id" ] }
        }
      }
    }
  }
}

将值与"$$lastViewed.members.groupId"返回的数组中的每个值进行比较，并返回true或false，具体取决于它是否匹配。

如果您无法使用$in，则只需将初始$filter更改为以下某项：

使用预告片$setIsSubset

   "cond": { 
     "$and": [
       { "$eq": ["$$lastViewed.userId", mongoose.Types.ObjectId(userId)] }, 
       { "$setIsSubset": [ ["$_id"], "$$lastViewed.members.groupId" ] }
     ]
   }

甚至$setIntersection和$size

   "cond": { 
     "$and": [
       { "$eq": ["$$lastViewed.userId", mongoose.Types.ObjectId(userId)] }, 
       { "$gt": [ 
         { "$size": { 
           "$setIntersection": [
             ["$_id"], 
             "$$lastViewed.members.groupId"
           ]
         }},
         0
       ] }
     ]
   }

“cross”导致“set”具有“多于”0个元素，表示该值存在于同一个引用的数组中。

请注意，在这些表单中，我们将"$_id"值设为["$_id"]，因为“集合比较”介于“集合”之间，而不是单个字段。

证明

这是您提供的文档更正文档：

{
    "_id" : ObjectId("5955ea4fd8099718330ab191"),
    "lastViewed" : [ 
        {
            "_id" : ObjectId("595218a7d346d27fb0bc1705"),
            "userId" : ObjectId("58c796d4344b9da4dfbe027b"),
            "members" : [ 
                {
                    "groupId" : ObjectId("5955ea4fd8099718330ab162"),
                    "lastViewedTime" : ISODate("2017-05-11T09:39:07.374Z")
                }, 
                {
                    "groupId" : ObjectId("5955ea4fd8099718330ab191"),
                    "lastViewedTime" : ISODate("2016-05-19T09:39:07.374Z")
                }
            ]
        }
    ]
}

以下是正在运行的管道阶段：

db.collection.aggregate([
  { "$project": {
    "viewedDetails": {
      "$map": {
        "input": {
          "$filter": {
            "input": "$lastViewed",
            "as": "lastViewed",
            "cond": { 
              "$and": [
                { "$eq": [ "$$lastViewed.userId", ObjectId("58c796d4344b9da4dfbe027b") ] }, 
                { "$in": [ "$_id", "$$lastViewed.members.groupId" ] }
              ]
            }
          }
        },
        "as": "v",
        "in": {
          "_id": "$$v._id",
          "userId": "$$v.userId",
          "members": { 
            "$filter": {
              "input": "$$v.members",
              "as": "m",
              "cond": { "$eq": [ "$$m.groupId", "$_id" ] }
            }
          }
        }
      }
    }
  }}
])

这是输出：

{
    "_id" : ObjectId("5955ea4fd8099718330ab191"),
    "viewedDetails" : [ 
        {
            "_id" : ObjectId("595218a7d346d27fb0bc1705"),
            "userId" : ObjectId("58c796d4344b9da4dfbe027b"),
            "members" : [ 
                {
                    "groupId" : ObjectId("5955ea4fd8099718330ab191"),
                    "lastViewedTime" : ISODate("2016-05-19T09:39:07.374Z")
                }
            ]
        }
    ]
}

如何$过滤条件与数组匹配的位置？

1 个答案:

证明