CODEWITHSUNDEEP
arraysmongodb
glneto
glneto

Reputation: 527

MongoDB - How to verify if a nested array item is contained in another nested array item on the same document?

The idea here is to return an array of documents of the users' followers with the information if this user is a friend of that follower or not.

So far I have:

db.getCollection('users').aggregate([
    { $match: { _id: ObjectId("588877d82523b4395039910a") } },
    { $lookup: {
            from: 'users',
            localField: 'followers',
            foreignField: '_id',
            as: 's_followers'
        }
    },
    { 
        $project: {
            "s_followers._id": 1,
            "s_followers.isFriend": {
                $in: ["s_followers.id",
                      { $setIntersection: ["$friends", "$followers"] }
                ]}
        }
    }
])

But the "s_followers.id" used in the $in operator doesn't seem to retrieve the _id information from the follower, so it always returns false.

When I use a ObjectId directly, I got the result I want:

"s_followers.isFriend": {
                    $in: [ObjectId("588877d82523b4395039910a"),
                          { $setIntersection: ["$friends", "$followers"] }
                    ]}

But I really need this ID to be a reference to the follower _id.

Expected result would be something like:

{
    "_id" : ObjectId("588877d82523b4395039910a"),
    "s_followers" : [ 
        {
            "_id" : ObjectId("5888687e56be8f172844d96f"),
            "isFriend" : true
        }, 
        {
            "_id" : ObjectId("5888ca27d79b8b03949a6e8c"),
            "isFriend" : false
        }
    ]
}

Thanks for your help!

UPD: A different approach (maybe easier), would be to use the ID of the user that I have (the one used on $match), but I would still need to get the reference for the follower's follower array

db.getCollection('users').aggregate([
    { $match: { _id: ObjectId("588877d82523b4395039910a") } },
    { $lookup: {
            from: 'users',
            localField: 'followers',
            foreignField: '_id',
            as: 's_followers'
        }
    }, {
        $project: {
            "firstName": 1,
            "s_followers._id": 1,
            "s_followers.firstName": 1,
            "s_followers.followers": 1,
            "s_followers.isFriend": { $in: [ObjectId("588877d82523b4395039910a"), "$s_followers.followers"] }
        }
    }
])

UPD2: The user data structure (the part that matters)

{
  followers: [{ type: mongoose.Schema.Types.ObjectId, ref: "User" }],
  friends: [{ type: mongoose.Schema.Types.ObjectId, ref: "User" }],
}

Upvotes: 2

Views: 58

Answers (2)

glneto
glneto

Reputation: 527

FOR VERSION 3.4.0+

Ok, just got it, I'll post here the code and my understanding of it:

db.getCollection('users').aggregate([
    { $match: { _id: ObjectId("588877d82523b4395039910a") } },
    { $lookup: {
            from: 'users',
            localField: 'followers',
            foreignField: '_id',
            as: 's_followers'
        }
    }, {
        $project: {
            "firstName": 1,
            "s_followers._id": 1,
            "s_followers.firstName": 1,
            "s_followers.followers": 1,
        }
    }, { 
        $unwind: "$s_followers"
    }, {
        $project: {
            "firstName": "$s_followers.firstName",
            "isFriend": { $in: [ObjectId("588877d82523b4395039910a"), "$s_followers.followers"] }
        }
    }
])

My understanding of it:

  1. $match: match the user I'm intended to get the followers of.
  2. $lookup: found each follower detail
  3. $project: select the information I want to return, also get the follower list of each follower
  4. $unwind: create a different document for each follower
  5. With the array unwinded, I can refer to the follower's followers array, and find my object id in it :)

Upvotes: 1

Shaishab Roy
Shaishab Roy

Reputation: 16805

In my example use followers friends list to check current user is friend or not. as {$arrayElemAt:["$s_followers.friends",0]} if want to find in followers then can use "$s_followers.followers"

You can try it. 

db.getCollection('user').aggregate([
    { $match: { _id: ObjectId("5714d190e6128b7e7f8d9008") } },
    {$unwind:"$followers"},
    { $lookup: {
            from: 'user',
            localField: 'followers',
            foreignField: '_id',
            as: 's_followers'
        }
    },
    {$project:{
        firstName:1,
        s_followers:{$arrayElemAt:["$s_followers",0]},
        isFriend:{$cond:[{
                $anyElementTrue:{
                  $map: {"input": {$arrayElemAt:["$s_followers.friends",0]},
                    "as": "el",
                    "in": { "$eq": [ "$$el", "$_id" ] }
                  }
                }
              },true,false]}
        } 
    },
    {$group:{
        _id:"$_id",
        s_followers:{$push:{_id:"$s_followers._id",isFriend:"$isFriend"}}
        }
    }
])

Upvotes: 0

Related Questions