Filter query on array of embedded documents(3 levels) in mongoDb - mongodb

Question

I have a collection of diagrams. Each diagram has several blocks. And each block is having several ports, which in itself has multiple objects as fields.
So far i am only able to apply $filter or $group up to level 1 on blocks. I am unable to filter up to ports based on some condition.
The document structure is: Diagram
[
{
"_id": "1",
"blocks": [
{
"port": [
{
"portType": {
"function": "input"
}
}
]
}
]
}
]
What i am trying to achieve is to get a list of all ports which have input as portType.function from the collection
db.collection.aggregate([{$project:{"blocks.ports":{$filter:{input:"$blocks.ports",as:"ports",cond:{$in:["input","$$ports.portType.function"]}}}}}]);
Expected output : list<Ports> which have only portType.function as input.
Actual output: Returns all the documents which have at least one port as "input" along with all the other ports as well.

Answer 1

If you want a list of Port objects you should use double $unwind stage, then $match for blocks.port.portType.function and finally extract only the ports using a $project stage
db.collection.aggregate([
{
"$unwind": "$blocks"
},
{
"$unwind": "$blocks.port"
},
{
"$match": {
"blocks.port.portType.function": "input"
}
},
{
$project: {
_id: 0,
portType: "$blocks.port.portType",
}
}
])
If your Port object contains other fields you want to extract, like for example Foo, you have just to add those fields in the $project stage like
$project: {
_id: 0,
portType: "$blocks.port.portType",
foo: "$blocks.port.foo"
...
}

Answer 2

you could simply use two $unwind stages to flatten the arrays, followed by a $match stage like this:
db.collection.aggregate([
{
"$unwind": "$blocks"
},
{
"$unwind": "$blocks.port"
},
{
"$match": {
"blocks.port.portType.function": "input"
}
}
])
result:
[
{
"_id": "1",
"blocks": {
"port": {
"portType": {
"function": "input"
}
}
}
}
]
try it online: mongoplayground.net/p/TOUCkCOSE7D