Aggregate + $project + $group + $sum + $count + $group again in MongoDB - node.js

Question

Consider the Schema :
const mongoose = require("mongoose");
const Schema = mongoose.Schema;
const EightWeekGamePlanSchema = new Schema({
Week: {
type: Number,
required: true
},
LeadId: {
type: String,
required: true
},
SupplierGeneralId: {
type: String,
required: true
},
// ... More properties
TotalClaimsLeftToBeClaimedByClientType: {
// incresed by 1 every time it's claimed
type: Number,
required: true
},
InsertDate: {
type: Date
// default: Date.now
}
});
module.exports = EightWeekGamePlan = mongoose.model(
"eightweekgameplan",
EightWeekGamePlanSchema
);
And consider the Mongo query :
EightWeekGamePlan.aggregate(
[
// 1.Group by supplier
// 2.Within each supplier group by month & year
// Use count & sum
{
$project: {
month: { $month: "$InsertDate" },
year: { $year: "$InsertDate" }
}
},
{
$group: {
_id: {
SupplierGeneralId: "$SupplierGeneralId",
month: "$month",
year: "$year"
},
ClaimsSummary : { $sum: "$TotalClaimsLeftToBeClaimedByClientType" } ,
// TotalLeadsPerSupplierAndDate : ...
// Here I want to group again , by LeadID and count all the
// unique LeadID's
}
}
]
I want to group by SupplierGeneralId and Month + Year of InsertDate ,
Summarize for each month the TotalClaimsLeftToBeClaimedByClientType
Group again but this time by the LeadID , and count all the unique LeadIds for each supplier (previously grouped by SupplierGeneralId, Month, Year).
However I'm getting
[ { _id: { month: 1, year: 2020 }, ClaimsSummary: 0 } ]
...even though there is data.
What's wrong with the pipeline ? and how can I group again to get the unique LeadIds for each supplier ?
Thanks
EDIT:
I've added more fields to the Project but now I'm getting empty array in the $push :
EightWeekGamePlan.aggregate(
[
// 1.Group by supplier
// 2.Within each supplier group by month & year
// Use count & sum
{ $sort: { SupplierGeneralId: 1 } },
{
$project: {
month: { $month: "$InsertDate" },
year: { $year: "$InsertDate" },
SupplierGeneralId: "$SupplierGeneralId",
TotalClaimsLeftToBeClaimedByClientType:
"$TotalClaimsLeftToBeClaimedByClientType"
}
},
{
$group: {
_id: {
SupplierGeneralId: "$SupplierGeneralId",
month: "$month",
year: "$year"
},
LeadsCollection: {
$push: {
LeadId: "$LeadId"
}
},
ClaimsSummary: { $sum: "$TotalClaimsLeftToBeClaimedByClientType" }
}
}
]
Output:
[
[0] {
[0] _id: {
[0] SupplierGeneralId: 'qCwHWFD1cBvrfPp5hdBL6M',
[0] month: 1,
[0] year: 2020
[0] },
[0] LeadsCollection: [
[0] {}, {}, {}, {}, {}, {}, {}, {},
[0] {}, {}, {}, {}, {}, {}, {}, {},
[0] {}, {}, {}, {}, {}, {}, {}, {},
[0] {}, {}, {}, {}, {}, {}, {}, {},
[0] {}, {}, {}, {}, {}, {}, {}, {},
[0] {}, {}, {}, {}, {}, {}, {}, {},
[0] {}, {}, {}, {}, {}, {}, {}, {},
[0] {}, {}, {}, {}, {}, {}, {}, {},
[0] {}, {}, {}, {}, {}, {}, {}, {}
[0] ],
[0] ClaimsSummary: 4
[0] }
[0] ]
SECOND EDIT :
EightWeekGamePlan.aggregate(
[
{ $sort: { SupplierGeneralId: 1 } },
{
$group: {
_id: {
SupplierGeneralId: "$SupplierGeneralId",
month: { $month: "$InsertDate" },
year: { $year: "$InsertDate" }
},
LeadsUploaded: {
$push: {
LeadId: "$LeadId"
}
},
Sells: { $sum: "$TotalClaimsLeftToBeClaimedByClientType" }
}
},
{
$project: {
Sells: "$Sells",
LeadsUploaded: {
$reduce: {
input: { $setUnion: "$LeadsUploaded.LeadId" },
initialValue: [],
in: {
$concatArrays: [
"$$value",
[
{
Lead_ID: "$$this"
}
]
]
}
}
}
}
}
]

Answer 1

You should just drop the $project stage, you're grouping right after so theres no real point of doing it, adding it just makes the pipeline less efficient.
Rewrite your pipeline as:
EightWeekGamePlan.aggregate(
[
// 1.Group by supplier
// 2.Within each supplier group by month & year
// Use count & sum
{ $sort: { SupplierGeneralId: 1 } },
{
$group: {
_id: {
SupplierGeneralId: "$SupplierGeneralId",
month: {"$month" : "$InsertDate"},
year: { $year: "$InsertDate" },
},
LeadsCollection: {
$push: {
LeadId: "$LeadId"
}
},
ClaimsSummary: { $sum: "$TotalClaimsLeftToBeClaimedByClientType" }
}
}
]