Question

Let’s say I have a collection:

[
    {
        _id: product_a,
        values: [
            { id: 1, value: 0 },
            { id: 2, value: 1 },
            { id: 3, value: 2 },
        ]
    },
    {
        _id: product_b,
        values: [
            { id: 1, value: 1 },
            { id: 2, value: 2 },
            { id: 3, value: 2 },
        ]
    },
    // etc ...
];

Is there any way to query this collection by aggregating the product of a sub-set of “values”?

If I query { values: [ 1, 3 ] }, I would get something like:

[
    {
        _id: product_a,
        result: 0 // since 0 * 2 = 0
    },
    {
        _id: product_b,
        result: 2 // since 1 * 2 = 2
    },
    // etc ...
];

Answer 1

Here is how you can do it with the schema the way you have it using 2.2 aggregation framework. Note that this would be a lot simpler if the id/value pairs were stored with id as key.

Your aggregation pipeline:

[
    {
        "$unwind" : "$values"
    },
    {
        "$match" : {
            "values.id" : {
                "$in" : [
                    1,
                    3
                ]
            }
        }
    },
    {
        "$project" : {
            "_id" : 1,
            "val1" : {
                "$cond" : [
                    {
                        "$eq" : [
                            "$values.id",
                            1
                        ]
                    },
                    "$values.value",
                    -1
                ]
            },
            "val3" : {
                "$cond" : [
                    {
                        "$eq" : [
                            "$values.id",
                            3
                        ]
                    },
                    "$values.value",
                    -1
                ]
            }
        }
    },
    {
        "$group" : {
            "_id" : "$_id",
            "val1" : {
                "$max" : "$val1"
            },
            "val3" : {
                "$max" : "$val3"
            }
        }
    },
    {
        "$project" : {
            "_id" : 1,
            "result" : {
                "$multiply" : [
                    "$val1",
                    "$val3"
                ]
            }
        }
    }
]

Note that the last step would do the job if the structure of your document was:
{_id: "product_x", values: [ {id1: value}, {id2: value} etc. ]}

Now run the query from the shell with:
> db.collection.aggregate(pipeline)

Or from your code via db.runCommand({"aggregate":"collection","pipeline":[...]})