Change type of field inside mongoDB aggregation and does $lookup utilises index on fields or not?

问题

I am performing joins in mongodb using $lookup, now i am facing a problem here. I have two collections first one to contains users all bookmarks brands and the second one contains all details about the brands.Now i am trying to return all brands details bookmarked by the user.

user_bookmarked Collection

{"mobile_no": "8971740148", "brands": ["5829c1df334d40e20e1d1c19", "5829c1df334d40e20e1d1c20", "5829c1df334d40e20e1d1c21"]}

brands Collection

{"_id": ObjectId("5829c1df334d40e20e1d1c19"), "brand_name": "Versace"}
{"_id": ObjectId("5829c1df334d40e20e1d1c20"), "brand_name": "Lee Cooper"}
{"_id": ObjectId("5829c1df334d40e20e1d1c21"), "brand_name": "Levis"}

My aggregation pipeline code is given below

                { $match: { mobile_no: mobile_no }},
                { $unwind: { path: "$brands", includeArrayIndex: "brandsposition"}},
                { $lookup: {from: "brands",localField: "brands",foreignField: "_id",as: "user_bookmarks"}},

Now the problem i am facing is that above code doesn't return anything as i am storing brands ids as string in my user_bookmarked collection but not as ObjectId so nothing is being returned. Now can anyone please tell me how can i change field type inside aggregation query.

Second thing i want to ask please tell me when using $lookup then does mongodb utilises index on foreign_field or not. Because i ran above aggregation pipeline with explain: true but i don't found any index that was utilised by above query i got this returned by the output.

db.user_bookmarked.runCommand('aggregate', {pipeline: [{ $match: { mobile_no: mobile_no }},
            { $unwind: { path: "$brands", includeArrayIndex: "brandsposition"}},
            { $lookup: {from: "brands",localField: "brands",foreignField: "_id",as: "user_bookmarks"}}], explain: true})
{
        "waitedMS" : NumberLong(0),
        "stages" : [
                {
                        "$cursor" : {
                                "query" : {
                                        "mobile_no" : "8971740148"
                                },
                                "queryPlanner" : {
                                        "plannerVersion" : 1,
                                        "namespace" : "test.restaurants",
                                        "indexFilterSet" : false,
                                        "parsedQuery" : {
                                                "mobile_no" : {
                                                        "$eq" : "8971740148"
                                                }
                                        },
                                        "winningPlan" : {
                                                "stage" : "COLLSCAN",
                                                "filter" : {
                                                        "mobile_no" : {
                                                                "$eq" : "8971740148"
                                                        }
                                                },
                                                "direction" : "forward"
                                        },
                                        "rejectedPlans" : [ ]
                                }
                        }
                },
                {
                        "$unwind" : {
                                "path" : "$brands",
                                "includeArrayIndex" : "brandsposition"
                        }
                },
                {
                        "$lookup" : {
                                "from" : "brands",
                                "as" : "user_bookmarks",
                                "localField" : "brands",
                                "foreignField" : "_id"
                        }
                }
        ],
        "ok" : 1
}

Now can anyone please help me out of here i have searched about both of this thing which is how to change type of field inside aggregation and does $lookup utilises indexes but Haven't found anything useful please help me out of here guys it would be really appreciable.

回答1:

You cannot convert the string to a object Id within the pipeline, you'll have to go though each document and convert it manually, using something like (you shouldnt be storing a mix match of types anyway, so it's probably worth updating in the long run):

how to convert string to numerical values in mongodb

as for does $lookup use index, If you look at the stats from this blog you'll see that indexes are used -

http://guyharrison.squarespace.com/blog/2016/7/4/join-performance-in-mongodb-32-using-lookup.html

回答2:

Try casting your brands to ObjectIds before your population:

user_bookmarked.brands.map((brand) => return mongoose.Types.ObjectId(brand) )

But you really should consider storing them as refs instead, your model should look something like:

const user_bookmarked = new mongoose.Schema({
    ...
    brands: [{type: mongoose.Schema.Types.ObjectId, ref: 'Brands'}],
    ...
    })

This way they will be ObjectIds from the start.!

Regarding the second question this post explains it I think: join-performance-in-mongodb-32-using-lookup

来源：https://stackoverflow.com/questions/41042482/change-type-of-field-inside-mongodb-aggregation-and-does-lookup-utilises-index