Sum of the values in nested dicts excluding some based on other key in MongoDB database

Viewed 54

I have a MongoDB database with the following format:

    {
    "_id": xxx,
    "timestamp": "1643649900000",
    "scores":
        [{
        "name": "APPL",
        "price": 80
        },
        {
        "name": "GOOGL",
        "price": 83,
        },
        {
        "name": "COMPI",
        "price": 76,
        },
        {
        "name": "and more names which also can change in the following documents",
        "price": 76,
        }]
    },
    {
    "_id": yyy,
    "time": "1644350400000",
    "scores":
        [{
        "name": "STCMP",
        "price": 33
        },
        {
        "name": "APPL",
        "price": 95,
        },
        {
        "name": "GOOGL",
        "price": 83,
        },
        {
        "name": "MINN",
        "price": 76,
        }]
    },

I need to sum all prices per time but excluding (or subtract from the sum) some.

My scores list has around 200 dicts and I want to exclude around 5 of them. I only managed to do the summing part, but after two days of search still can't manage to exclude with my limited knowledge.

toBeExcluded = ["APPL", "GOOGL"]
sums.aggregate([
            {
                "$unwind" : "$scores"
            },
            {
                "$group": {
                    "_id": "$time",
                    "total": {
                        "$sum": "$scores.price"
                    }
                }
            },
            {
                "$addFields":{
                    "timeAdj": {"$toInt": [{"$subtract":[{"$divide": ["$_id", 1000]}, 300]}]}
                }
            },
            {
                "$sort": {"timeAdj":1}
            }
            ]))
1 Answers

use $cond and check names to be excluded using $in, in your $sum:

mongo playground

db.collection.aggregate([
  {
    "$unwind": "$scores"
  },
  {
    "$group": {
      "_id": "$time",
      "total": {
        "$sum": {
          "$cond": [
            {
              "$in": [
                "$scores.name",
                [
                  "APPL",
                  "GOOGL"
                ]
              ]
            },
            0,
            "$scores.price"
          ],
          
        }
      }
    }
  },
  
])
Related