Nested array aggregation query in mongodb with projected fileds?

Question

I have the following data. I want to use mongodb aggregate function to extract suppliers and customers from the data shown below . I want to project specific fields and remove rest of the fields from output data.

 {
        "_id" : ObjectId("577f9a25bea25d480d8f1895"),
        "password" : "12345",
        "mobile" : "9582223889",
        "email" : "a@b.com",
        "name" : "ashiush jindal",
        "invoice" : [{
            "name" : "Ashish Jindal",
            "dname" : "jindalbe",
            "email" : "a@b.com",
            "password" : "12345",
            "role" : ["customer"]
          }],
        "inventory" : [{ }, {
            "item" : "Levis Jeans",
            "sku" : "12345",
            "buy" : [{
                "bp" : "jindalbe",
                "qty" : 50,
                "created" : "9/7/2016"
              }]
          }, {
            "item" : "Levis Trouser",
            "sku" : "123",
            "selling_price" : 2000,
            "buy" : [{
                "bp" : "jindalbe",
                "qty" : 90,
                "created" : "9/7/2016",
                "price_per_qty" : 1000
              }, {
                "bp" : "jindalbe",
                "qty" : 60,
                "created" : "9/7/2016",
                "price_per_qty" : 1000
              }, {
                "bp" : "jindalbe",
                "qty" : 60,
                "created" : "9/7/2016",
                "price_per_qty" : 1000
              }, {
                "bp" : "jindalbe",
                "qty" : 60,
                "created" : "9/7/2016",
                "price_per_qty" : 1000
              }, {
                "bp" : "jindalbe",
                "qty" : 60,
                "created" : "9/7/2016",
                "price_per_qty" : 5000
              }, null]
          }],
        "business_partner" : [{
            "name" : "Ashish Jindal",
            "dname" : "jindalbe",
            "email" : "a@b.com",
            "password" : "12345",
            "role" : ["customer"]
          }, {
            "name" : "Ashish Jindal",
            "dname" : "jindalbe",
            "email" : "a@b.com",
            "password" : "12345",
            "role" : ["customer"]
          }, {
            "name" : "Ashish Kumar",
            "dname" : "jindal",
            "email" : "a@b.com",
            "password" : "12345",
            "role" : ["supplier"]
          }],
        "__v" : 0
      }

I want to extract all the business partners name who are suppliers also .

In the form show below .

{
   suppliers:["Ashish Kumar"]
}

I want to do it using aggregation query .

The solution I tried is .

          [
                { $match: { "_id": ObjectId(req.headers["token"]) } },{
                    $project: {
                        suppliers: {
                            $filter: {
                                input: '$business_partner.role',
                                as: 'role',
                                cond: {"$eq":["$$bp.role","supplier"]}
                            }
                        },
                        _id: 0
                    }
                }
            ];

Answer 1

You can $unwind your arrays and use $match conditions. Then use $addToSet to add all your supplier to a single array. There will be no supplier duplicate for each _id :

db.device.aggregate([{
  "$unwind": "$business_partner"
}, {
  "$unwind": "$business_partner.role"
}, {
  "$match": {
    "_id": ObjectId("577f9a25bea25d480d8f1895"),
    "business_partner.role": "supplier"
  }
}, {
  "$project": {
    "_id": 1,
    "business_partner": 1
  }
},{
    "$group": {
        _id:"$_id",
        supplier: {
            $addToSet: '$business_partner'
        }
    }
}]);

This is compatible with mongo 2.6 (where there is no $filter )

It will give you this kind of output :

{
  "_id": ObjectId("577f9a25bea25d480d8f1895"),
  "supplier": [{
    "name": "Ashish Kumar",
    "dname": "jindal",
    "email": "a@b.com",
    "password": "12345",
    "role": "supplier"
  }]
}