Skip to content

Views

Virtual views are aggregation pipelines stored in MongoDB that act as collections for reading operations. You can use the View class the same way as Document for find and aggregate operations.

Here are some examples.

Create a view:

from pydantic import Field

from beanie import Document, View


class Bike(Document):
    type: str
    frame_size: int
    is_new: bool


class Metrics(View):
    type: str = Field(alias="_id")
    number: int
    new: int

    class Settings:
        source = Bike
        pipeline = [
            {
                "$group": {
                    "_id": "$type",
                    "number": {"$sum": 1},
                    "new": {"$sum": {"$cond": ["$is_new", 1, 0]}}
                }
            },
        ]

Initialize Beanie:

from motor.motor_asyncio import AsyncIOMotorClient

from beanie import init_beanie


async def main():
    uri = "mongodb://beanie:beanie@localhost:27017"
    client = AsyncIOMotorClient(uri)
    db = client.bikes

    await init_beanie(
        database=db, 
        document_models=[Bike, Metrics],
        recreate_views=True,
    )

Create bikes:

await Bike(type="Mountain", frame_size=54, is_new=True).insert()
await Bike(type="Mountain", frame_size=60, is_new=False).insert()
await Bike(type="Road", frame_size=52, is_new=True).insert()
await Bike(type="Road", frame_size=54, is_new=True).insert()
await Bike(type="Road", frame_size=58, is_new=False).insert()

Find metrics for type == "Road"

results = await Metrics.find(Metrics.type == "Road").to_list()
print(results)

>> [Metrics(type='Road', number=3, new=2)]

Aggregate over metrics to get the count of all the new bikes:

results = await Metrics.aggregate([{
    "$group": {
        "_id": None,
        "new_total": {"$sum": "$new"}
    }
}]).to_list()

print(results)

>> [{'_id': None, 'new_total': 3}]

A better result can be achieved by using find query aggregation syntactic sugar:

results = await Metrics.all().sum(Metrics.new)

print(results)

>> 3