basic mongodb nested sets search, only want to retrieve single value - mongodb

Question

I just started with document based datastores around 3-4 hours and I have a basic problem that I want to understand.
{
"_id": "5527e5ae06e55c02049bd114",
"owner": "John Doe",
"customers" : ["5527e3c806e55c01dad3a132", "5527e3c806e55c01dad3a133", "5527e3c806e55c01dad3a134"],
"location" : [
{
"address": "Wall St",
"location_id": "123123213",
"vendor" : [
{
"name" : "hello 123",
"price" : "3",
"serial" : "000000009730978e"
},
{
"name" : "hello abc",
"price" : "3.5",
"serial" : "0000000097308888"
}
]
},
{
"address" : "PCH 1",
"location_id": "987987",
"vendor" : [
{
"name" : "hello 456342",
"price" : "4",
"serial" : "000000009733452435"
},
{
"name" : "hello sdfsdg",
"price" : "4.5",
"serial" : "0000000095243532453"
}
]
}
]
}
So how can I find location.serial.price?
db.test.find_one( {"location.location_id" : "123123213" , "location.vendor.serial" : "000000009730978e" } )
would returns the entire object but I am just interested in location.serial.price where these conditions match.
Thanks a lot,
Ben

Answer 1

Usually you would use the positional-operator ($) to refer to array entries. But unfortunately this operator has a serious limitation: it does not work with nested arrays. So it does not help you in this case.
What you can do instead is use an aggregation pipeline which unwinds both arrays and then matches the serial.
db.test.aggregate([
// create a stream of location-documents
{ $unwind: "$location" },
// filter the stream by location-id
{ $match: { "location.id" : "123123213" },
// expand the remaining stream further to individual vendor-documents
{ $unwind: "$vendor" },
// filter the stream by serial
{ $match: { "location.vendor.serial": "000000009730978e" } }
]);
Keep in mind that aggregation can become quite slow. It also has a limitation of 16MB per aggregation step. You can avoid that limit with the allowDiskUse:true option, but that makes it even slower. So when you have lots of data and performance is a concern, you might want to reconsider your database schema.

Answer 2

Mongodb aggregation use here, below query will satisfied your criteria
db.collectionName.aggregate({
"$unwind": "$location"
},
{
"$match": {
"location.location_id": "123123213"
}
},
{
"$unwind": "$location.vendor"
},
{
"$match": {
"location.vendor.serial": "000000009730978e"
}
},
{
"$project": {
"serial": "$location.vendor.serial",
"price": "$location.vendor.price",
"_id": 0
}
}).pretty()