Performance improvements for BSON cursor

Working with Data Developer Tools
,
1

Hi there,
I’m using mongocxx driver.
My application creates about 10+ documents per second into a collection. Like realtime data. Every now and then I need to find a huge number of those documents (let say 25k). At the end I need one JSON array of all the docs. The most time consuming part is the iteration over the resulting cursor and the parsing to JSON. Here we are talking about a consumption range of second(s).
Any ideas this can be approved?

This is basically the structure:

{
    "_id": {
        "$oid": "6329bc45f6740000d5002413"
    },
    "time": {
        "$numberDouble": "570974.7541653"
    },
    "data": {    }
}

The query is mostly about time.

Yes, projection improves the parsing time proportional to the individual size of each found document. Is there potential in aggregations?

Thanks