Is it possible to join one field of an array unwind onto the unwound array?

user_964823 0
Viewed 179

Pretty new to mongo and haven't been able to figure out how to perform a query.

I have an accounts collection that looks like this:

{
    "_id" : ObjectId("1"),
    "time" : ISODate("2018-10-20T05:57:15.372Z"),
    "profileId" : "1",
    "totalUSD" : "1015.5513030613",
    "accounts" : [
        {
            "_id" : ObjectId("2"),
            "accountId" : "1",
            "currency" : "USD",
            "balance" : "530.7934159683763000",
            "available" : "530.7934159683763",
            "hold" : "0.0000000000000000",
            "exchangeRateUSD" : "1"
        },
        {
            "_id" : ObjectId("5"),
            "accountId" : "4",
            "currency" : "BTC",
            "balance" : "0.0759214200000000",
            "available" : "0.07592142",
            "hold" : "0.0000000000000000",
            "exchangeRateUSD" : "6384.995"
        },
    ],
}

I store only exchangeRateUSD for each currency, and not exchangeRateXXX where XXX is currency name, because there can be an arbitrary number of currencies and currency pairs. But when I query the accounts collection it will always be queried by a currency pair, eg: BTC-USD. Keeping it simple for now, I can assume the currency pair will always be XXX-USD.

When I query the accounts collection I'd like to add a 'virtual' field to each account object: exchangeRateCrypto and then on the top-level accounts document I'd like to add totalCrypto which would just be the total account value in the given crypto. Eg: USD account balance * exchangeRateCrypto + crypto account balance * exchangeRateCrypto (which would equal 1).

My current query without the exchangeRateCrypto and totalCrypto looks like:

db.accounts.aggregate([
  { $unwind: '$accounts' },
  { $match: { 'accounts.currency': { $in: [ 'USD', 'BTC' ] }}},
  {
    $group: {
      _id: '$_id',
      time: { $first: '$time' },
      profileId: { $first: '$profileId' },
      accounts:  { $push: '$accounts' },
      totalUSD: { $sum: { $multiply: [ { $toDouble: '$accounts.balance' }, { $toDouble: '$accounts.exchangeRateUSD' } ] } }
    }
  }
]);

I'm trying to figure out how to 'reach' into the BTC row and calculate the exchangeRateCrypto by simply doing 1 / exchangeRateUSD and then projecting/returning the accounts document and subdocument as:

{
    "_id" : ObjectId("1"),
    "time" : ISODate("2018-10-20T05:57:15.372Z"),
    "profileId" : "1",
    "totalUSD" : "1015.5513030613",
    "totalCrypto" : "0.1590527953",   // 530.7934159683763 * 0.0001566171939 + 0.07592142 * 1
    "accounts" : [
        {
            "_id" : ObjectId("2"),
            "accountId" : "1",
            "currency" : "USD",
            "balance" : "530.7934159683763000",
            "available" : "530.7934159683763",
            "hold" : "0.0000000000000000",
            "exchangeRateUSD" : "1",
            "exchangeRateCrypto" : "0.0001566171939",   //  1 / 6384.995
        },
        {
            "_id" : ObjectId("5"),
            "accountId" : "4",
            "currency" : "BTC",
            "balance" : "0.0759214200000000",
            "available" : "0.07592142",
            "hold" : "0.0000000000000000",
            "exchangeRateUSD" : "6384.995",
            "exchangeRateCrypto" : "1"
        },
    ],
}

but haven't been able to figure out a good way of doing this.

It seems it should be pretty straightforward, but still learning Mongo.

Any tips?

Thanks!

aggregation-framework mongodb-query mongodb
1 Answers

The solution might be a bit long and probably it can be shortened however I want you to understand proposed way of thinking step by step.

var secondCurrency = "BTC";
var secondCurrencyFieldName = "exchangeRate" + secondCurrency;
var secondCurrencyFieldNameRef = "$" + secondCurrencyFieldName;
var totalFieldName = "total" + secondCurrency;

db.accounts.aggregate([
    { $unwind: "$accounts" },
    { $match: { "accounts.currency": { $in: [ "USD", secondCurrency ] }}},
    {
        $group: {
            _id: "$_id",
            time: { $first: "$time" },
            profileId: { $first: "$profileId" },
            accounts:  { $push: "$accounts" },
            totalUSD: { $sum: { $multiply: [ { $toDouble: "$accounts.balance" }, { $toDouble: "$accounts.exchangeRateUSD" } ] } }
        }
},
{
    $addFields: {
        [secondCurrencyFieldName]: {
            $filter: {
                input: "$accounts",
                as: "account",
                cond: { $eq: [  "$$account.currency", secondCurrency ] }
            }
        }
    }
},
{
    $addFields: {
        [secondCurrencyFieldName]: {
            $let: {
                vars: { first: { $arrayElemAt: [ secondCurrencyFieldNameRef, 0 ] } },
                in: { $toDouble: "$$first.exchangeRateUSD" }
            }
        }
    }
},
{
    $addFields: {
        accounts: {
            $map: {
                input: "$accounts",
                as: "account",
                in: {
                    $mergeObjects: [
                        "$$account",
                            { 
                            [secondCurrencyFieldName]: {
                                $cond: [ { $eq: [ "$$account.currency", secondCurrency ] }, 1, { $divide: [ 1, secondCurrencyFieldNameRef ] } ]
                                } 
                            }
                    ]
                }
            }
        }
    }
},
{
    $addFields: {
        [totalFieldName]: {
            $reduce: {
                input: "$accounts",
                initialValue: 0,
                in: {
                    $add: [
                        "$$value",
                        { $multiply: [ { $toDouble: "$$this.balance" }, "$$this." + secondCurrencyFieldName ] }
                    ]
                }
            }
        }
    }
}
]).pretty()

So we can start with $addFields which can either add new field to existing document or repace existing field. After the $group stage you have to find the USD-XXX exchange rate (using $filter and $let + $arrayElemAt in the next pipeline stage). Having this value you can use $addFields again combined with $map and $mergeObjects to add new field to nested array and that field will represent the ratio between USD and XXX currency. Then you can use $addFields again with $reduce to get the total of all accounts for XXX currency.

Output:

{
    "_id" : ObjectId("5beeec9fef99bb86541abf7f"),
    "time" : ISODate("2018-10-20T05:57:15.372Z"),
    "profileId" : "1",
    "accounts" : [
            {
                    "_id" : ObjectId("5beeec9fef99bb86541abf7d"),
                    "accountId" : "1",
                    "currency" : "USD",
                    "balance" : "530.7934159683763000",
                    "available" : "530.7934159683763",
                    "hold" : "0.0000000000000000",
                    "exchangeRateUSD" : "1",
                    "exchangeRateBTC" : 0.00015661719390539853
            },
            {
                    "_id" : ObjectId("5beeec9fef99bb86541abf7e"),
                    "accountId" : "4",
                    "currency" : "BTC",
                    "balance" : "0.0759214200000000",
                    "available" : "0.07592142",
                    "hold" : "0.0000000000000000",
                    "exchangeRateUSD" : "6384.995",
                    "exchangeRateBTC" : 1
            }
    ],
    "totalUSD" : 1015.5513030612763,
    "exchangeRateBTC" : 6384.995,
    "totalexchangeRateBTC" : 0.15905279535242806
}
user_62389770
Related