How to properly filter Django Queryset and apply annotate?

Question

Django

I want to filter chats with the participation of the user who requests these chats and count the number of participants for each chat

class Chat(models.Model):
    users = models.ManyToManyField(User, related_name='chats')


def get_queryset(self):
    return Chat.objects.annotate(user_count=Count('users')).filter(users__id=self.request.user.id)


def get_queryset(self):
    return Chat.objects.filter(users__id=self.request.user.id).annotate(user_count=Count('users'))

The first method returns such data. They are correct and I need them:

{
        "id": 146,
        "user_count": 2
    },
    {
        "id": 109,
        "user_count": 3
    },
    {
        "id": 22,
        "user_count": 11
    }

The second returns this:

{
        "id": 146,
        "user_count": 1
    },
    {
        "id": 109,
        "user_count": 1
    },
    {
        "id": 22,
        "user_count": 1
    }

As I understand it, the second method is better, since I first check for the presence of a user in "users", collect such chats and then count the participants as it.
In the first case, I first count the participants for all chats (although I don’t need all chats), and then filter by the presence of a user.

Can you tell me why the second method counts incorrectly?
Is there any difference in speed between these two methods?

Answer 1

an error in the second because at first you cut off all other users (except one) with a filter and count the number of users for it, the
difference in speed will be if the number of participants is large.
try like this:

def get_queryset(self):
    qs = Chat.objects.filter(users__id=self.request.user.id).values_list('id')
    return Chat.objects.filter(id__in=qs).annotate(user_count=Count('users'))