Understanding the changes in MongoDB new C# driver (Async and Await)

SexyMF picture SexyMF · May 7, 2015 · Viewed 11.9k times · Source

The new C# driver is totally Async and in my understanding twists a little bit the old design patterns such as DAL in n-tier architecture.

In my Mongo DALs I use to do:

public T Insert(T entity){
     _collection.Insert(entity);
     return entity;
}

This way I can get the persisted ObjectId.

Today, everything is Async such as InsertOneAsync.
How would Insert method will now return the entity when the InsertOneAsync will be done? Can you show an example?

Answer

mnemosyn picture mnemosyn · May 7, 2015

It's helpful to understand the basics of async / await because it's a somewhat leaky abstraction and has a number of pitfalls.

Essentially, you have two options:

  • Remain synchronous. In this case, it's safe to use .Result and .Wait() on the async calls, respectively, e.g. something like

    // Insert:
    collection.InsertOneAsync(user).Wait();
    
    // FindAll:
    var first = collection.Find(p => true).ToListAsync().Result.FirstOrDefault();
    
  • Go async in your code base. Doing it async is quite 'infectious', unfortunately, so either you convert pretty much everything to async, or not. Careful, mixing sync and async incorrectly will lead to deadlocks. Using async has a number of advantages, because your code can continue to run while MongoDB is still working, e.g.

    // FindAll:
    var task = collection.Find(p => true).ToListAsync();
    // ...do something else that takes time, be it CPU or I/O bound
    // in parallel to the running request. If there's nothing else to 
    // do, you just freed up a thread that can be used to serve another 
    // customer...
    // once you need the results from mongo:
    var list = await task;