Mongodb $lookup joins all collection instead of matching object - mongodb

Question

so I am trying to do a $lookup with Mongodb but I have a strange output.
I have two collections, "sites" and "consumptions".
sites :
{
"_id" : ObjectId("5b26db6e7f59e825909da106"),
"siteId" : 49,
"industry" : "Commercial Property",
"sub_industry" : "Shopping Center/Shopping Mall",
"square_feet" : 497092,
"latitude" : 41.2161756,
"longitude" : -78.14809154,
"timezone" : "America/New_York",
"timezone_offset" : "-04:00",
"__v" : 0
}
consumptions :
{
"_id" : ObjectId("5b26db907f59e825909f3d2a"),
"timestamp" : 1325382000,
"dttm_utc" : ISODate("2012-01-01T00:40:00Z"),
"value" : 2.8956,
"estimated" : 0,
"anomaly" : "",
"site" : [
{
"_id" : ObjectId("5b26db727f59e825909da16a")
}
],
"__v" : 0
}
This is the $lookup I am trying to do :
db.consumptions.aggregate([
{
$lookup:
{
from: "sites",
localField: "site.id",
foreignField: "id",
as: "site"
}
}
])
The expected output would be to have the detail of the site in each consumption :
{
"_id" : ObjectId("5b26db907f59e825909f3d2a"),
"timestamp" : 1325382000,
"dttm_utc" : ISODate("2012-01-01T00:40:00Z"),
"value" : 2.8956,
"estimated" : 0,
"anomaly" : "",
"site" : [
{
"_id" : ObjectId("5b26db6e7f59e825909da106"),
"siteId" : 49,
"industry" : "Commercial Property",
"sub_industry" : "Shopping Center/Shopping Mall",
"square_feet" : 497092,
"latitude" : 41.2161756,
"longitude" : -78.14809154,
"timezone" : "America/New_York",
"timezone_offset" : "-04:00",
"__v" : 0
}
],
"__v" : 0
}
This is the output I am getting with the $lookup :
{
"_id" : ObjectId("5b26db907f59e825909f3d2a"),
"timestamp" : 1325382000,
"dttm_utc" : ISODate("2012-01-01T00:40:00Z"),
"value" : 2.8956,
"estimated" : 0,
"anomaly" : "",
"site" : [
{
"_id" : ObjectId("5b26db6e7f59e825909da0f3"),
"siteId" : 6,
"industry" : "Commercial Property",
"sub_industry" : "Shopping Center/Shopping Mall",
"square_feet" : 161532,
"latitude" : 34.78300117,
"longitude" : -106.8952497,
"timezone" : "America/Denver",
"timezone_offset" : "-06:00",
"__v" : 0
},
{
"_id" : ObjectId("5b26db6e7f59e825909da0f4"),
"siteId" : 8,
"industry" : "Commercial Property",
"sub_industry" : "Shopping Center/Shopping Mall",
"square_feet" : 823966,
"latitude" : 40.32024733,
"longitude" : -76.40494239,
"timezone" : "America/New_York",
"timezone_offset" : "-04:00",
"__v" : 0
}, ... (all the sites details are listed)
],
"__v" : 0
}
Thank you in advance for your help !

Answer 1

You need to first $unwind the site array to match site._id to the foreign field _id and then $group to rolling back into the arrays again.
db.collection.aggregate([
{ "$unwind": "$site" },
{ "$lookup": {
"from": Site.collection.name,
"localField": "site._id",
"foreignField": "_id",
"as": "site"
}},
{ "$unwind": "$site" },
{ "$group": {
"_id": "$_id",
"value": { "$first": "$value" },
"estimated": { "$first": "$estimated" },
"anomaly": { "$first": "$anomaly" },
"timestamp": { "$first": "$timestamp" },
"dttm_utc": { "$first": "$dttm_utc" },
"site": { "$push": "$site" }
}}
])
And if you have mongodb 3.6 then you can try this
db.collection.aggregate([
{ "$unwind": "$site" },
{ "$lookup": {
"from": Site.collection.name,
"let": { "siteId": "$site._id" },
"pipeline": [
{ "$match": { "$expr": { "$eq": [ "$_id", "$$siteId" ] } } }
],
"as": "site"
}},
{ "$unwind": "$site" },
{ "$group": {
"_id": "$_id",
"value": { "$first": "$value" },
"estimated": { "$first": "$estimated" },
"anomaly": { "$first": "$anomaly" },
"timestamp": { "$first": "$timestamp" },
"dttm_utc": { "$first": "$dttm_utc" },
"site": { "$push": "$site" }
}}
])
Make sure you should put Site.collection.name correctly

Answer 2

I think that The $lookup doesn't work directly with an array.
try using $unwind first.