mongoosejs: populating an array of objectId's from different schemas

Question

I'm testing out an idea and have made some contrived examples for discussion.

I have the following schema's models:

var outerSchema = new Schema({
    title: String,
    refD: [{type: Schema.ObjectId, ref: 'InnerOne'},
           {type: Schema.ObjectId, ref: 'InnerTwo'}]
});

var innerOneSchema = new Schema({
    title: String
});

var innerTwoSchema = new Schema({
    turtle:String
});

var Outer = mongoose.model('Outer', outerSchema);
var InnerOne = mongoose.model('InnerOne', innerOneSchema);
var InnerTwo = mongoose.model('InnerTwo', innerTwoSchema);

I have successfully pushed an example of each of these to Mongo. This has resulted in three collections with a refD array that includes the ObjectId's of the innerOne and innerTwo objects, looking good from mongo's point of view:

{
    "_id" : ObjectId("56625b7c752995f505637864"),
    "title" : "LOOK AT MY CHILDREN!",
    "refD" : [
        ObjectId("56625b7c752995f505637862"),
        ObjectId("56625b7c752995f505637863")
    ],
    "__v" : 1
}

However when I use the populate command as so:

Outer
    .findOne({title: 'LOOK AT MY CHILDREN!'})
    .populate('refD')
    .exec(function(err,res){
        if(err) console.log(err);

        console.log(res);
    });

the result is as follows:

{ refD: 
   [ { __v: 0,
       title: 'Im the first type inside!',
       _id: 56625b7c752995f505637862 } ],
  __v: 1,
  title: 'LOOK AT MY CHILDREN!',
  _id: 56625b7c752995f505637864 }

So structurally this is correct however it looks as though the populate function has only populated the first of the two types.

Am I pushing the limits of what can be performed or is there a way to handle the population of different objects?

I had considered using discriminators to make ALL models children of a single parent class but I would prefer not to, if I do that I may as well use a relational database.

------EDIT------ At this point it doesn't seem like this is doable without modifying or creating a plug in.

For anyone who was thinking the same thing I decided for now I would go back to a more traditional approach, I've created a relatedRecordSchema that stores an Id and a record type and some other useful metadata. This essentially pushes the logic for population up to the app.

In my use case this isn't so bad as I don't always want to populate all of the various related records but for the application or user to decide on the fly.

This isn't an answer so I will leave the question open hoping someone takes the initiative and builds a plugin ;)

Cheers B

Answer 1

To my knowledge populate works for one Schema. You could try to run the populate two times to the 'refD' by giving different model parameter.

More information about populate: Mongoose populate

I didn't test this through, but possible way to solve this problem:

Outer
.findOne({title: 'LOOK AT MY CHILDREN!'})
.populate({
  path: 'refD',
  model: 'InnerOne'
})
.populate({
  path: 'refD',
  model: 'InnerTwo'
})
.exec(function(err,res){
    if(err) console.log(err);

    console.log(res);
});