Still cannot find the correct way to populate fields in the callback mapreduce

I know it’s a new feature, since 3.6. But i have tried everything.

I have a model declared like this

var TagsMapReduce = new Schema({
    _id: {type: Schema.ObjectId, ref: 'Tag'},
    value: {type : Number}
});

mongoose.model('TagsMapReduce', TagsMapReduce);

i have set in my mapreduce object

o.out = { replace: 'TagsMapReduce' }

then I have a mapreduce

PostEntry
  .mapReduce(o, function (err, results) {
    //console.log(results);
    results
    .find()
    //.populate({ path: '_id', model: 'TagsMapReduce', select:'_id name' })
    .exec(function(err, res){
        console.log(res);
//output is [ { _id: 5182be986ae2bec804000006, value: 1 }, { _id: 5182bea86ae2bec804000007, value: 1 } ]

    })
  });

if i uncomment the populate, the result are now showing in place of ids null

[ { _id: null, value: 1 }, { _id: null, value: 1 } ]

and are not populated. I have double checked the name of the model, and the name of the referring schema, everything seems fine.

say tags are

var TagSchema = new Schema({
    name: {type : String, default : '', trim : true, unique: true}
});

mongoose.model('Tag', TagSchema);

[talamaska]

I think I have found the problem. when populating

.populate({ path: '_id', model: 'TagsMapReduce', select:'_id name' })

I should define the model to which I want to create the reference, not the model that I want to use as structure for getting the reference.

so if I put

.populate({ path: '_id', model: 'Tag', select:'_id name' })

Mongoose: tags.find({ _id: { '$in': [ ObjectId("517e7149167f889018000005"), ObjectId("517e714f167f889018000006
"), ObjectId("517e7155167f889018000007"), ObjectId("517e7165167f889018000009") ] } }) { fields: { name: 1, _id: 1 }, safe: undefined, lean: true }

//populated tags
[ { _id: { _id: 517e7149167f889018000005, name: 'Buy and Sell' }, value: 2 },
  { _id: { _id: 517e714f167f889018000006, name: 'Fun' }, value: 1 },
  { _id: { _id: 517e7155167f889018000007, name: 'Tutorial' }, value: 1 },
  { _id: { _id: 517e7165167f889018000009, name: 'Important' }, value: 1 } ]

which is exactly what I want. I will make a reformat so that in the end that json looks like this

[ { _id: 517e7149167f889018000005, name: 'Buy and Sell' , count: 2 },
  { _id: 517e714f167f889018000006, name: 'Fun' , count: 1 },
  { _id: 517e7155167f889018000007, name: 'Tutorial' , count: 1 },
  { _id: 517e7165167f889018000009, name: 'Important' , count: 1 } ]

Also I have found that, I have to define a schema for replace with dbref in it, for every mapreduce on every different collection and that i have to define the same model name when populating, it is not working automatically. if i don’t define the dbref in the mapreduce models, and i put the model name just in the populate, nothing happens, I don’t see in the console any request to mongodb showing it is getting the id and the name from another collection. This all seems a non-intentional.
I think if i define a modelname to be used for the population, it should not require a db ref in the model and it should make the ref automatically. that way if you have 10 mapreduces to 10 collection which have same format _id, value, you could define only one model. I guess if you have more than one ObjectID in the mapreduce resulted model, and you want to populate them all , you could use the options passed to populate - path, model, select to create a reference to another collection.

It’s very likely that my thinking is not correct, I’m relatively new in mongodb. So I would appreciate any feedback on the question.

[markotom]

I use this map for replace values of _id after map/reduce

docs = docs.map(function(doc, key){
  var d   = doc._id;
  d.value = doc.value;
  return d;
});