MongoDB将[{k1：v1}，{k2：v2}]转换为{v1：v2}

Question

I have a collection with the following schema:

// ----- SOURCE SCHEMA ----- //

{
    "_id" : ObjectId("5564e8937c32c85f03ea028a"),
    "sku" : "MD01"
},
{
    "_id" : ObjectId("5564e8c97c32c85f03ea028b"),
    "sku" : "MD02"
},
{
    "_id" : ObjectId("5564e8dd7c32c85f03ea028c"),
    "sku" : "MD03"
}

// ----- SOURCE SCHEMA ----- //

Now I want to transform these documents into the following format:

// ----- DESIRED FORMAT ----- //

{
    "5564e8dd7c32c85f03ea028c": "MD01",
    "5564e8c97c32c85f03ea028b": "MD02",
    "5564e8dd7c32c85f03ea028c": "MD03"
}

// ----- DESIRED FORMAT ----- //

Currently I have to use two for loop in server side javascript for this task, but I think it will face performance problems if the collection going bigger.

So how can I archive this by using only Mongodb queries ? Is the performance of this approach better?

Answer 1

Using Map-Reduce you can achieve the desired output:

// map each object to { "XXXXXXXX": sku }
map = function() {
  obj = {}
  obj[this._id.valueOf()] = this.sku
  emit(null, obj)
}

// Aggregate all objects in one (very?) big one
// by copying individual properties
reduce = function(key, values) {
  result = {}
  for (idx in values) {
    for (k in values[idx]) {
      result[k] = values[idx][k]
    }
  }

  return result
}

Producing:

> db.test.mapReduce(map, reduce, {out: {inline:1}}).results
[
    {
        "_id" : null,
        "value" : {
            "5564e8937c32c85f03ea028a" : "MD01",
            "5564e8c97c32c85f03ea028b" : "MD02",
            "5564e8dd7c32c85f03ea028c" : "MD03"
        }
    }
]

---

That being said, this is not efficient and will eventually hit the 16M limit as you are in fact requesting an aggregation of all your DB objects in one document only.