How to design a collection to store friendships?

Working with Data
1

If you have a Users collection and there is a many to many relationship, such as subscriber and subscribed, then you have 2 fields to hold these 2 distinct participants of this relationship.

However, there is no particular relationship between the 2 friends, how do you name the fields? It’s not predictable in what field to look for a particular member?

2

Hi @Big_Cat_Public_Safety_Act ,

Not sure what exactly your question is?

Usually many to many will be stored as reference arrays inside the related documents:

// Profiles
{ 
_id : "A",
name : "aaa",
"subscribedFriends" : [ { "_id" : "B", name : "bbb" } , { "_id" : "C", name : "ccc"}],
"followingFriends" : [{"_id" : "D", name : "ddd" }]
}

{ 
_id : "B",
name : "bbb",
"subscribedFriends" : [ { "_id" : "D", name : "ddd" } , { "_id" : "C", name : "ccc"}]
"followingFriends" :  [ { "_id" : "A", name : "aaa" } ]
}
...

To avoid outliers where users have thousands of followers or subscribers we use the outlier pattern:

Does that answer your question?

Thanks,
Pavel

3

The question is in the event of the outlier where the data needs to be stored in a separate collection. And also, there is no distinction between the members of the relationship. In your example, you have subscribedFriends and followingFriends. In my case, it’s just friends. Both members are just friends.

4

Ok so they will apear in each others lists:


{ 
_id : "A",
name : "aaa",
"friends" : [ { "_id" : "B", name : "bbb" } , { "_id" : "C", name : "ccc"} ... 500 friends],
bucketFriends : 500
}

{ 
_id : "A-outlier-1",
name : "aaa",
"friends" : [ { "_id" : "X" ... 400 friends],
bucketFriends : 400
}

{ 
_id : "B",
name : "bbb",
"friends" :  [ { "_id" : "A", name : "aaa" } ]
}

Here the outlier sit in the same collection so _id : “A” , and its outlier : “A-outlier-a” are still in the same collection to get all friends of “A” run:

db.collection.find({ _id : /^A/ })

Thanks
Pavel