Retrieve all values of a collection in mongodb

Question

I'm looking to retrieve all values of my collection in a single set :

Example :

` "_id" : {
        "origin" : "xx",
        "destination" : "yy"
    },
    "paths" : [ 
        [ 
            "tt"
        ]
    ]
}

/* 2 */
{
    "_id" : {
        "origin" : "aa",
        "destination" : "bb"
    },
    "paths" : [ 
        [ 
            "cc"
        ]
    ]

}

/* 3 */
{
    "_id" : {
        "origin" : "xy",
        "destination" : "yx"
    },
    "paths" : [ 
        [ 
            "yy", 
            "tt", 
            "cc"
        ]
    ]

}`

Expected output :

Nodes : {"xx", "yy", "aa", "bb","xy", "yx"}

I tried $setUnion but it doesn't work because I have string

$addToset

is impossible to add both fields : "origin" and "destination"

How to retrieve all values of my collection's fields (id.origin and id.destination) to a set?

Thank you

Answer 1

Since this is very special case where you want custom formatting, your best bet will be map-reduce feature of MongoDB. However, this formatting can be achieved with help of aggregation framework too. I'm adding both solutions.

Aggregation framework:

db.collection.aggregate([
  {
    $group:{
      _id:null, 
      origin:{
        $addToSet:"$_id.origin"
      }, 
      destination:{
        $addToSet:"$_id.destination"
      }
  }},
  {
    $project:{
      _id:0, 
      Nodes:{
        $setUnion:["$origin","$destination"]
      }
  }}
])

Output:

{ 
    "Nodes" : [
        "yy", 
        "yx", 
        "xx", 
        "bb", 
        "aa", 
        "xy"
    ]
}

Map Reduce:

db.collection.mapReduce(
    function () {
        emit(1, this._id);
    },
    function (key, values) {
        var o = {};
        o.Nodes = [];

        for (var i = 0; i < values.length; i++) {
            o.Nodes.push(values[i].origin);
            o.Nodes.push(values[i].destination);
        }

        return o;
    },
    {
        out: { inline: 1 }
    });

Output:

{ 
    "results" : [
        {
            "_id" : NumberInt(1), 
            "value" : {
                "Nodes" : [
                    "xx", 
                    "yy", 
                    "aa", 
                    "bb", 
                    "xy", 
                    "yx"
                ]
            }
        }
    ], 
    "timeMillis" : NumberInt(22), 
    "counts" : {
        "input" : NumberInt(3), 
        "emit" : NumberInt(3), 
        "reduce" : NumberInt(1), 
        "output" : NumberInt(1)
    }, 
    "ok" : NumberInt(1)
}

results.values.Nodes contains your desired result.

Answer 2

In the aggregation pipeline first you can have two sets (Originset and DestinationSet) and after that you can use setUnion to have the union set of the two set.