Reputation: 1055
$lookup is used to perform a left outer join to an unsharded collection in the same database to filter in documents from the “joined” collection for processing in Mongo.
{
$lookup:
{
from: <collection to join>,
localField: <field from the input documents>,
foreignField: <field from the documents of the "from" collection>,
as: <output array field>
}
}
Could the foreignField
be the field of the nested document of from
collection?
For example, there are two collections as following.
history
collection
[{
id:'001',
history:'today worked',
child_id:'ch001'
},{
id:'002',
history:'working',
child_id:'ch004'
},{
id:'003',
history:'now working'
child_id:'ch009'
}],
childsgroup
collection
[{
groupid:'g001', name:'group1'
childs:[{
id:'ch001',
info:{name:'a'}
},{
id:'ch002',
info:{name:'a'}
}]
},{
groupid:'g002', name:'group1'
childs:[{
id:'ch004',
info:{name:'a'}
},{
id:'ch009',
info:{name:'a'}
}]
}]
so, this aggregation
code could be executed like this?
db.history.aggregate([
{
$lookup:
{
from: "childsgroup",
localField: "child_id",
foreignField: "childs.$.id",
as: "childinfo"
}
}
])
So I want to get the result like this.
[{
id:'001',
history:'today worked',
child_id:'ch001',
childinfo:{
id:'001',
history:'today worked',
child_id:'ch001'
}
}, .... ]
Isn't this possible?
Upvotes: 2
Views: 957
Reputation: 49945
There's no positional operator for $lookup but you can use custom pipeline
in MongoDB 3.6 to define custom join conditions:
db.history.aggregate([
{
$lookup: {
from: "childsgroup",
let: { child_id: "$child_id" },
pipeline: [
{ $match: { $expr: { $in: [ "$$child_id", "$childs.id" ] } } },
{ $unwind: "$childs" },
{ $match: { $expr: { $eq: [ "$childs.id", "$$child_id" ] } } },
{ $replaceRoot: { newRoot: "$childs" } }
],
as: "childInfo"
}
}
])
First $match
added to improve performance: we want to find only those documents from childsgroup
that contain matching child_id
and then we can match subdocuments after $unwind
stage.
Upvotes: 3