MongoDB Correct Schema for aggregated data

Question

I have a big collection that holds lots of stats, since I want to generate reports, I am running a daily cron which aggregates data from the main collection to a smaller one

Answer 1

Personally, seems it looks like these are "delivery dates" for parts of an order I would do this:

{
    'order_id': 'LaOPX',
    'user_id': '777',
    'parts': [
        { "date": ISODate("2015-04-18T00:00:00Z"), "qty": 100 },
        { "date": ISODate("2015-04-19T00:00:00Z"), "qty": 20 }
    ]
}

Where the dates where "actual date objects" in the database . If you wanted everything for all of user "777" data in all records then you can do:

db.collection.aggregate([
    // Match the user between dates
    { "$match": { 
        "user_id": "777", 
        "parts.date": { 
            "$gte": new Date("2015-04-18"), "$lt": new Date("2015-04-20")
        }
    }},

    // Unwind the array entries
    { "$unwind": "$parts" },

    // Filter the required dates
    { "$match": { 
        "parts.date": { 
            "$gte": new Date("2015-04-18"), "$lt": new Date("2015-04-20")
        }
    }},

    // Group per user
    { "$group": {
        "_id": "$user_id",
        "total": { "$sum": "$parts.qty" }
    }}
])

It's much more flexible to use real dates in the data as range queries will always work as they should