MongoDb query needs to be framed - mongodb

Question

I have a collection named 'Dealer' with following fields.
1) userId
2) dealerId
3) code
4) origin (having values 'UUM', 'MMT', 'TTC')
In the above collection, I need to get unique dealer records with unique combination of (userId + dealerId). If two records are having same 'userId' and 'dealerId' then I need to check the origin field and the record with value not equal to 'UUM' needs to be returned in resultset.
This is what I reached up to, as of now:
db.dealer.aggregate(
[
{
$sort: { origin: 1 }
},
{
$group:
{
_id: {
dealerId:"$dealerId",
recordId:"$recordId"
},
origin: {
$first: "$origin"
}
}
}
])
How can I make an aggregation query for the above scenario.Any help is appreciated.
Thanks in advance.

Answer 1

To reduce memory usage, I would filter the docs that have don't have a UUM origin.
Then the group by dealers and users.
I'm not sure why you're sorting and grouping by dealer + record.
Setup
> db.dealerTest.find()
{ "_id" : 1, "userId" : "u1", "dealerId" : "d1", "code" : 1, "origin" : "UUM" }
{ "_id" : 2, "userId" : "u1", "dealerId" : "d1", "code" : 2, "origin" : "UUM" }
{ "_id" : 3, "userId" : "u1", "dealerId" : "d1", "code" : 3, "origin" : "TTC" }
{ "_id" : 4, "userId" : "u2", "dealerId" : "d1", "code" : 4, "origin" : "TTC" }
{ "_id" : 5, "userId" : "u2", "dealerId" : "d1", "code" : 5, "origin" : "MMT" }
{ "_id" : 6, "userId" : "u2", "dealerId" : "d2", "code" : 6, "origin" : "UUM" }
Code:
db.dealerTest.aggregate([{
$match: {
origin: {
$ne: "UUM"
}
}
}, {
$group: {
_id: {
dealerId: "$dealerId",
userId: "$userId"
},
docIds: { $addToSet : "$_id" },
origins: { $addToSet : "$origin" }
}
}]);
Output:
{ "_id" : { "dealerId" : "d1", "userId" : "u2" }, "docIds" : [ 5, 4 ], "origins" : [ "MMT", "TTC" ] }
{ "_id" : { "dealerId" : "d1", "userId" : "u1" }, "docIds" : [ 3 ], "origins" : [ "TTC" ] }