MongoDB $lookup on nested document

Question

I'm new to mongo and struggling mightily with the following. I have 2 collections structured as below. For the life of me, I can't figure out how to do a $lookup on the school collection. Reading other posts, I'm definitely using ObjectId for the reference as well as the foreign field.

Below is my structure:

Alumni:

{
    "_id": "john",
    "items": [
        {
            "name": "John",
            "items": [
                {
                    "school": ObjectId("56de35ab520fc05b2fa3d5e4"),
                    "grad": true
                },
                {
                    "school": ObjectId("56de35ab520fc05b2fa00000"),
                    "grad": false
                }
            ]
        },
        {
            "name": "Johnny"
            // notice no nested items, this doc should still be included in result
        },
        {
            "name": "Jon",
            "items": [
                {
                    "school": ObjectId("56de35ab520fc05b2fa11111"),
                    "grad": false
                }
            ]
        }
     ]
}

Schools

{
    _id: ObjectId("56de35ab520fc05b2fa3d5e4"),
    name: "Some University",
    street: "ABC Boulevard"
}

What I'm looking to get:

{
    "_id": "john",
    "items": [
        {
            "name": "John",
            "items": [
                {
                    "school": ObjectId("56de35ab520fc05b2fa3d5e4"),
                    "grad": true,
                    "schoolInfo":     {
                        _id: ObjectId("56de35ab520fc05b2fa3d5e4"),
                        name: "Some University",
                        street: "ABC Boulevard"
                    }
                },
                {
                    "school": ObjectId("56de35ab520fc05b2fa00000"),
                    "grad": true,
                    "schoolInfo":     {
                        _id: ObjectId("56de35ab520fc05b2fa00000"),
                        name: "Another University",
                        street: "123 Boulevard"
                    }
                }
            ]
        },
        {
            name: "Johnny"
        },
        {
            "name": "Jon",
            "items": [
                {
                    "school": ObjectId("56de35ab520fc05b2fa11111"),
                    "grad": true,
                    "schoolInfo":     {
                        _id: ObjectId("56de35ab520fc05b2fa11111"),
                        name: "Some University",
                        street: "ABC Boulevard"
                    }
                }
            ]
         }
     ]
}

The query I've tried to no avail:

db.alumni.aggregate([
      {$match: {_id: 'john'}}, 
      {$lookup: {
                from: 'schools', 
                localField: 'items.items.school', 
                foreignField: '_id', 
                as: 'schoolInfo'}}
 ])

Any help would be greatly appreciated!

Answer 1

in this case there is required a nice play with $unwind and $project in aggregation framework

please see below:

db.alumni.aggregate([
    {$match: {_id: 'john'}},
    {$unwind:"$items"},
    {$unwind:"$items.items"},
    {$lookup: {
        from: 'schools', 
        localField: 'items.items.school', 
        foreignField: '_id', 
        as: 'schoolInfo'}},
    {$unwind:"$schoolInfo"},
    {$project:{
        "_id":1,
        "items":[{
            "name":"$items.name",
            "items":[{
            "school":"$schoolInfo._id"    ,
            "grad":"$items.items.grad"    ,
            "schoolInfo":"$schoolInfo"
            }]
        }]            
    }}
]).pretty()

to see how it works - try removing aggregation stages from query and check document structure.

Answer 2

Would that handle better the multiple schools in the $items.items array?

db.alumni.aggregate([
    {$match: {_id: 'john'}},
    {$unwind:"$items"},
    {$unwind:"$items.items"},
    {$lookup: {
        from: 'schools', 
        localField: 'items.items.school', 
        foreignField: '_id', 
        as: 'schoolInfo'}},
    {$unwind:"$schoolInfo"},
    {$group:{
        _id: {
            _id: '$_id',
            name: '$items.name',
        },
        items: {
            $push: {
                'grad': '$items.items.grad',
                'school': '$schoolInfo._id'
                'schoolInfo': '$schoolInfo'
            }
        }
    }},
    {$group:{
        _id: '$_id._id',
        items: {
            $push: {
                'name': '$_id.name',
                'items': '$items'
            }
        }
    }}
]).pretty()

I didn't address the case for missing $items.items but you can look at $unwind empty array Additionally, it'd be better to leave an empty array instead of nothing when there's no entry, ie have

{
    name: "Johnny",
    items: [],
},