Get sorted Firestore documents when denormalizing data - swift

Question

I am using Firestore for my app on which users can publish Posts, stored in the posts collection:
posts
{postID}
content = ...
attachementUrl = ...
authorID = {userID}
Each User will also have a timeline in which the posts of the people they follow will appear. For that I am also keeping a user_timelines collection that gets populated via a Cloud Function:
user_timelines
{userID}
posts
{documentID}
postID = {postID}
addedDate = ...
Because the data is denormalized, if I want to iterate through a user's timeline I need to perform an additional (inner) query to get the complete Post object via its {postID}, like so:
db.collection("user_timelines").document(userID).collection("posts")
.orderBy("addedDate", "desc").limit(100).getDocuments() { (querySnap, _) in
for queryDoc in querySnap.documents {
let postID = queryDoc.data()["postID"] as! String
db.collection("posts").document("postID").getDocument() { (snap, _) in
if let postDoc = snap {
let post = Post(document: postDoc)
posts.append(post)
}
}
}
}
The problem is by doing so I am loosing the order of my collection because we are not guaranteed that all the inner queries will complete in the same order. I need to keep the order of my collection as this will match the order of the timeline.
If I had all the complete Post objects in the timeline collections there would be not issue and the .orderBy("addedDate", "desc").limit(100) would work just fine keeping the Posts sorted, but If I denormalize I cant seem to find a correct solution.
How can I iterate through a user's timeline and make sure to get all the Post objects sorted by addedDate even when denormalizing data?
I was thinking of creating a mapping dictionary postID/addedDate when reading the postIDs, and then sort the Post at the end using this dictionary, but I am thinking there must be a better solution for that?
I was expecting this to be a common issue when denormalizing data, but unfortunately I couldnot find any results. Maybe there's something I am missing here.
Thank you for your help!

Answer 1

What you can do is enumerate the loop where you perform the inner query, which simply numbers each iteration. From there, you could expand the Post model to include this value n and then sort the array by n when you're done.
db.collection("user_timelines").document(userID).collection("posts").orderBy("addedDate", "desc").limit(100).getDocuments() { (querySnap, _) in
for (n, queryDoc) in querySnap.documents.enumerated() {
let postID = queryDoc.data()["postID"] as! String
db.collection("posts").document("postID").getDocument() { (snap, _) in
if let postDoc = snap {
let post = Post(document: postDoc, n: n)
posts.append(post)
}
}
}
posts.sort(by: { $0.n < $1.n })
}
The example above actually won't work because the loop is asynchronous which means the array will sort before all of the downloads have completed. For that, consider using a Dispatch Group to coordinate this task.
db.collection("user_timelines").document(userID).collection("posts").orderBy("addedDate", "desc").limit(100).getDocuments() { (querySnap, _) in
let dispatch = DispatchGroup()
for (n, queryDoc) in querySnap.documents.enumerated() {
dispatch.enter() // enter on each iteration
let postID = queryDoc.data()["postID"] as! String
db.collection("posts").document("postID").getDocument() { (snap, _) in
if let postDoc = snap {
let post = Post(document: postDoc, n: n)
posts.append(post)
}
dispatch.leave() // leave no matter success or failure
}
}
dispatch.notify(queue: .main) { // completion
posts.sort(by: { $0.n < $1.n })
}
}