How to do multiple joins between two collection in mongodb using lambda function?

Question

I have two collections 1) user_posts 2)user_profile. find the below collection data for your reference.

1) user_posts collection

_id :ObjectId("5d519f861c9d4400005ebd1b")
userid : ObjectId("5d518caed55bc00001d235c1")
media : "hello.jpg"
type : "jpg"
created : " "
modified : " "
like : Array
       0 : Object
           userid : ObjectId("5d518caed55bc00001d235c1")
           status : "like"
       1 : Object
           userid : ObjectId("5d518da6d55bc00001d235c2")
           status : "happy"
comment : Array
       0 : Object
           userid : ObjectId("5d518caed55bc00001d235c1")
           comment : "hello"
       1 : Object
           userid : ObjectId("5d518da6d55bc00001d235c2")
           comment : "welcome"
share : Array
       0 : Object
           userid : ObjectId("5d518caed55bc00001d235c1")
           status : "shared"
       1 : Object
           userid : ObjectId("5d518da6d55bc00001d235c2")
           status : "shared"

2) User_profile collection

 _id : ObjectId("5d518caed55bc00001d235c1")
 username : "ramesh",
 photo :  " ",
 created : " ",
 modified : " "

 _id : ObjectId("5d518da6d55bc00001d235c2")
 username : "shekar",
 photo :  " ",
 created : " ",
 modified : " "

Now i tried to get the profile details from user_profile in lambda function. but i didn't get the details. find the below lambda function code.

def lambda_handler(event, context):

print("Received event: " + json.dumps(event, indent=1))

user_posts = db.user_posts

Userid = event['userid']
uid = ObjectId(Userid)

dispost = list(user_posts.aggregate([{
"$match" : { "userid" : uid }
},
{ "$graphLookup" : 
     {
       "from" : "user_profile",
       "startWith" : "$like.userid",
       "connectFromField" : "like.userid",
       "connectToField" : "_id",
       "as" : "userdetails"
     }
},
{ "$graphLookup" : 
     {
       "from" : "user_profile",
       "startWith" : "$comment.userid",
       "connectFromField" : "comment.userid",
       "connectToField" : "_id",
       "as" : "userdetails1"
     }
}
{ "$graphLookup" : 
     {
       "from" : "user_profile",
       "startWith" : "$share.userid",
       "connectFromField" : "share.userid",
       "connectToField" : "_id",
       "as" : "userdetails2"
     }
}
]))     
disair = json.dumps(dispost, default=json_util.default)
return json.loads(disair)

but i didn't get the output. i need output like this below.

_id :ObjectId("5d519f861c9d4400005ebd1b")
userid : ObjectId("5d518caed55bc00001d235c1")
username : "ramesh"
photo : " ",
media : "hello.jpg"
type : "jpg"
created : " "
modified : " "
like : Array
       0 : Object
           userid : ObjectId("5d518caed55bc00001d235c1")
           status : "like"
           username : "ramesh"
           photo : " "
       1 : Object
           username : "shekar"
           photo : " "
           userid : ObjectId("5d518da6d55bc00001d235c2")
           status : "happy"
           username : "shekar"
           photo : " "
comment : Array
       0 : Object
           userid : ObjectId("5d518caed55bc00001d235c1")
           comment : "hello"
           username : "ramesh"
           photo : " "
       1 : Object
           userid : ObjectId("5d518da6d55bc00001d235c2")
           comment : "welocme"
           username : "shekar"
           photo : " "
share : Array
       0 : Object
           userid : ObjectId("5d518caed55bc00001d235c1")
           status : "shared"
           username : "ramesh"
           photo : " "
       1 : Object
           userid : ObjectId("5d518da6d55bc00001d235c2")
           status : "shared"
           username : "shekar"
           photo : " "

can you please help me the solutions. Thanks in advance.

Ramesh Reddy · Accepted Answer · 2019-08-21 15:38:32Z

Please check this:

db.collection("user_posts").aggregate(
{ $match: {"userid" : uid}},
{ $unwind: '$like' },
{ $lookup: { from: "users", localField: "like.userid", foreignField: "_id", as: 
"users" }},
{ $group: {
    _id: "$_id",
    like: { $push: { $mergeObjects: ['$like', { $arrayElemAt: [ "$users", 0 ] } ]}},
    data: { $first: "$$ROOT" }
}},
{ $replaceRoot: { newRoot: { $mergeObjects: ['$data', { like: "$like"} ]} } },
{ $unwind: '$comment' },
{ $lookup: { from: "users", localField: "comment.userid", foreignField: "_id", as: 
 "users" }},
{ $group: {
    _id: "$_id",
        comment: { $push: { $mergeObjects: ['$comment', { $arrayElemAt: [ "$users", 0 
 ] } ]}},
        data: { $first: "$$ROOT" }
}},
{ $replaceRoot: { newRoot: { $mergeObjects: ['$data', { comment: "$comment"} ]} } },
{ $unwind: '$share' },
{ $lookup: { from: "users", localField: "share.userid", foreignField: "_id", as: 
"users" }},
{ $group: {
    _id: "$_id",
    share: { $push: { $mergeObjects: ['$share', { $arrayElemAt: [ "$users", 0 ] } 
]}},
    data: { $first: "$$ROOT" }
}},
{ $replaceRoot: { newRoot: { $mergeObjects: ['$data', { share: "$share"} ]} } },
{ $project: { users: 0 }}
)

you will get output, changes add/remove fields name in project aggregation as per your requirement

whoami - fakeFaceTrueSoul · Accepted Answer · 2019-08-21 15:55:53Z

0

Please check this :

    db.user_posts.aggregate([{ $match: { _id: ObjectId("5d519f861c9d4400005ebd1b") } }, {
    $lookup:
    {
        from: "user_profile",
        let: { userIdToBeCompared: "$userid", like: '$like', comment: '$comment', share: '$share' },
        pipeline: [
            {
                $match:
                {
                    $expr:
                    {

                        $or:
                            [
                                { $eq: ["$_id", "$$userIdToBeCompared"] },
                                { $in: ["$_id", "$$like.userid"] },
                                { $in: ["$_id", "$$comment.userid"] },
                                { $in: ["$_id", "$$share.userid"] }
                            ]
                    }
                }
            }

        ],
        as: "data"
    }
}])

Ok there are way too many things has to happen in your requirement and with your current data structure it might not be an easy thing & a good idea to implement all your needs from your DB layer, So over here final response data will have all the mappings for your needs, get it to your code and fetch the needed details to each section based on userid, as of what I've tried I thought that would be ideal scenario to implement !!

edited Aug 21, 2019 at 15:55

answered Aug 15, 2019 at 5:30

whoami - fakeFaceTrueSoul

18.1k6 gold badges36 silver badges51 bronze badges

5 Comments

whoami - fakeFaceTrueSoul Over a year ago

@Ramesh : Please try this and refactor with your findings and feel free to add things, I can take a look to enhance it further !!

Ramesh Reddy Over a year ago

bro i get the output but it is not similar to my expected output. output like this, all user_posts userid's matching with user_profile collection _id's and get matched user_profile _id's details display in separate data array. I tried to used $project aggregation to display the username, photo fileds in like, comment, share arrays but it is not working

whoami - fakeFaceTrueSoul Over a year ago

@RameshReddy : Yes exactly iterate thru all the fields and fetch the needed details based on matching keys is not that easy in this case thru mongoDB, Since it could be easily done by code can you try it at your end, meanwhile I would certainly work on it & will update if any feasible solution is achieved..

Ramesh Reddy Over a year ago

Bro any other solutions according my expected output.

whoami - fakeFaceTrueSoul Over a year ago

@RameshReddy : nope your desired o/p has to be achieved thru code, doing it thru DB is tedious, I've tried couple of ways someway or the other getting stuck i.e; too many thing here, Or you can've multiple $facet stages for each and merge the results in the code, that as well doesn't serve your requirement in DB query.

Collectives™ on Stack Overflow

How to do multiple joins between two collection in mongodb using lambda function?

2 Answers 2

Comments

5 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

5 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related