[MongoHook] support creation of time series collections

[gemunet]

Description

Mongo creates the collections automatically, but from mongo v5 you can create time series collections, for this you can create them with the 'db.createCollection' method that is not exposed in the MongoHook api.

https://www.mongodb.com/docs/manual/core/timeseries/timeseries-procedures/

Could support be added by exposing the createCollection api or by adding time series creation arguments to the current get_collection ?

Use case/motivation

support new time series collections

Related issues

No response

Are you willing to submit a PR?

• Yes I am willing to submit a PR!

Code of Conduct

• I agree to follow this project's Code of Conduct

[boring-cyborg]

Thanks for opening your first issue here! Be sure to follow the issue template! If you are willing to raise PR to address this issue please do so, no need to wait for approval.

[shunnNet]

@gemunet
The MongoHook use pymongo under the ground, it already support time series collection

You can do it by following

      mongo_hook = MongoHook(conn_id="mongo_default")
      mongo_conn = mongo_hook.get_conn()
      mongo_db = mongo_conn["db_name"]

     # create time collection here
      mongo_db.create_collection("time_test", timeseries= {
          "timeField": "timestamp",
          "metaField": "metadata",
          "granularity": "hours"
      })

      # use the collection
      mongo_collection = mongo_db["time_test"]

      # add document
      mongo_collection.insert_one(
        {
          "metadata": { "sensorId": 5578, "type": "temperature" },
          "timestamp": dt.utcnow(),
          "temp": 12
        }
      )

the mongo_db variable actually pymongo Database object
you can check pymongo doc

[nailo2c]

Hi, I'd like to give this one a try.