CODEWITHSUNDEEP
mongodbdatabase-designdatabase-schema
Damian
Damian

Reputation: 974

Facebook "like" feature with mongodb schema/endpoint design

I came up with multiple solutions but none of them seems to be "future proof".

Let me iterate my solutions first.

Acceptance criteria of the feature

  1. User can like or unlike a post
  2. User can see "filled heart" on the post that he liked already

Assumption

There is User collection and Post collection

Possible Solutions

1. Put array of user references who liked the post.

Problem

Super simple, but once "post" grows too much, all users that query the post will be affected. Also putting potentially infinite array in a document.

2. Put array of Post references to User

Problem

Same with above, but only the user with growing data may be affected.

3. Like schema with reference to User and Post and add counter to Post

Problem

This document will grow very rapidly, many users will like many posts everyday. How do I query which posts a user liked? What if this collection grows to millions of like documents?

So far I have not found the right answer, at first, I thought #1 was way to go, then quickly changed my mind to #3, then I was like - what if it grows fast and becomes painful to query to find "likes" from millions of documents every time I fetch posts? Now I am thinking #2 is "safer" option but I may be wrong. If I go #3, I'd like to separate /likes and/posts endpoints, so that's another issue. - how will I query if a user liked from list of posts? Should it be separate singular endpoint? But it will be too many calls each time list of posts loads up. I am confused :(

Thanks for reading.

Upvotes: 1

Views: 1006

Answers (1)

GitGitBoom
GitGitBoom

Reputation: 1912

Door #3! Then your gonna use a unique compound index with User+Post to prevent duplicates and to have fast query times.

When you query posts your going to want to use an aggregate pipe so you can $lookup to see if the 'like' document exists for each post.

https://docs.mongodb.com/manual/reference/operator/aggregation/lookup/

https://docs.mongodb.com/manual/core/index-compound/index.html

Should get you started.

Upvotes: 2

Related Questions