MongoDB $lookup on array of objects

Question

Categories

    {
        "_id" : ObjectId("61740086893f048528d166b9"),
        "name": "Category1",
        "tracks" : [ 
            "61c65353565a2d9a1cd3020d", 
            "61c74518962dc3efb96c3438", 
            "61c74775703176a6f72df444"
        ]
    }

Tracks

    {
        "_id" : ObjectId("61c65353565a2d9a1cd3020d"),
        "name" : "Track1",
        "categoryId" : ObjectId("61740086893f048528d166b9"),
        "creatorId" : ObjectId("61c6478304e98ed63e8ee7d3"),
        "thumbnailId" : ObjectId("61c65353565a2d9a1cd3020c"),
        "plays" : [],
        "media" : {
            "type" : "wav",
            "url" : ""
        },
        "status" : "approved",
        "downloads" : [],
        "uploadedDate" : 1640387411
    }

Assuming that I have 5 categories and each category has many tracks ID, I wanna get N last tracks for each category so I used this code below

    categories.aggregate([
        {
            $project: {
                tracks: { $slice: ["$tracks", -2] },
            },
        },
    ]

And the response is

    [
        {
            "_id": "61740086893f048528d166b9",
            "tracks": [
                "61c74518962dc3efb96c3438",
                "61c74775703176a6f72df444"
            ]
        },
        {
            "_id": "61740094893f048528d166c1",
            "tracks": []
        },
        {
            "_id": "617400a0893f048528d166cb",
            "tracks": []
        }
    ]

So far it's good, but the question is how can I replace each category's tracks from an array of IDs to an array of objects?

I tried $loopup but I probably didn't implement the localField correctly.

Expected result

    [
        {
            "_id": "61740086893f048528d166b9",
            "tracks": [
                {
                    "_id": ObjectId("61c74518962dc3efb96c3438")
                    ...
                },
                {
                    "_id": ObjectId("61c74775703176a6f72df444")
                    ...
                }
            ]
        },
        {
            "_id": "61740094893f048528d166c1",
            "tracks": []
        },
        {
            "_id": "617400a0893f048528d166cb",
            "tracks": []
        }
    ]

***** UPDATE *****

I'm trying to replace the creatorId by createdBy which is an object of the users from the users collection

Users

    {
      "_id": ObjectId("61c6478304e98ed63e8ee7cb"),
      "email": "[email protected]",
      "username": "USER999",
      "tracks": [
        ObjectId("61c65353565a2d9a1cd3020d"),
      ],
    }

The expected result should be

    [
        {
            "_id": "61740086893f048528d166b9",
            "tracks": [
                {
                    "_id": ObjectId("61c74518962dc3efb96c3438"),
                    "createdBy": {
                        "_id": "userId"
                        ...
                    },
                    ...
                },
                {
                    "_id": ObjectId("61c74775703176a6f72df444"),
                    "createdBy": {
                        "_id": "userId"
                        ...
                    }
                    ...
                }
            ]
        },
        {
            "_id": "61740094893f048528d166c1",
            "tracks": []
        },
        {
            "_id": "617400a0893f048528d166cb",
            "tracks": []
        }
    ]

In addition to the solution below by ray, I added the code here https://mongoplayground.net/p/8AjmnL-vhtz

The createdBy is at the top level but not under every track

ray · Accepted Answer · 2022-01-15 15:33:44Z

1

$lookup is the correct way for you to find the corresponding object in Tracks collection. Why your code does not work is that you are storing strings in tracks array in Categories collection; while the _id of Tracks collection is ObjectId. There will be no $lookup result as the datatypes do not match. What you can do is converting the strings to ObjectId by using $toObjectId in a $map, and then do the $lookup

db.categories.aggregate([
  {
    $project: {
      tracks: {
        $slice: [
          "$tracks",
          -2
        ]
      }
    }
  },
  {
    $project: {
      tracks: {
        "$map": {
          "input": "$tracks",
          "as": "t",
          "in": {
            "$toObjectId": "$$t"
          }
        }
      }
    }
  },
  {
    "$lookup": {
      "from": "tracks",
      let: {
        t: "$tracks"
      },
      pipeline: [
        {
          $match: {
            $expr: {
              "$in": [
                "$_id",
                "$$t"
              ]
            }
          }
        }
      ],
      "as": "tracks"
    }
  }
])

Here is the Mongo playground for your reference.

answered Jan 15, 2022 at 15:33

ray

15.6k12 gold badges97 silver badges105 bronze badges

Sign up to request clarification or add additional context in comments.

5 Comments

daniel93 Over a year ago

thanks for your help, I have one more question, each track has creatorId, how can i replace the creatorId by createdBy which is an object with the user's info? I have a collection of users

ray Over a year ago

@daniel93 the idea is the same. Simply perform a $lookup to the users collection. You should be fine this time as the createdBy is already objectId type. If you need help with a concrete example, please update your question with the new sample dataset and expected output.

daniel93 Over a year ago

updated with mongodb playground link

ray Over a year ago

@daniel93 try to do the $lookup in the subpipeline when $lookup tracks, like this

daniel93 Over a year ago

that's right, I looked where to perform the nested lookup tried couple of things but I missed the pipeline... was confused but I learnt something new! :)

Collectives™ on Stack Overflow

MongoDB $lookup on array of objects

1 Answer 1

5 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

5 Comments

Your Answer

Sign up or log in

Post as a guest

Related