DJango REST框架：如何使用ModelViewSet编写创建，列表，更新和删除

Question

I'm still new to python/django and currently I'm trying to add REST API using django rest framework. This is currently what I have:

urls.py

actions_url = {'actions': {'get': 'list', 'post': 'create'}}
actions_obj = {'actions': {'get': 'retrieve', 'delete': 'destroy', 'put': 'update'}}

urlpatterns = [
   path('book', BookViewSet.as_view(**actions_url)),
   path('book/<str:isbn>', BookViewSet.as_view(**actions_obj))
]

models.py (The model doesn't have PK)

class Book(models.Model):
   BNAME = models.CharField(
       db_column='BNAME',
       max_length=30,
       default='')

   BISBN = models.CharField(
       db_column='BISBN',
       max_length=10,
       default='')

views.py

class BookSerializer(serializers.ModelSerializer):
   name = serializers.CharField(source="BNAME")
   isbn = serializers.CharField(source="BISBN")

       class Meta:
           model = models.Book
           fields = ('name', 'isbn')


class BookViewSet(viewsets.ModelViewSet)
   queryset = Book.objects.all()
   serializer_class = BookSerializer

   def list(self, request):
       return queryset

   def create(self, request):
       pass

   def retrieve(self, request, *args, **kwargs):
       pass

   def update(self, request, *args, **kwargs):
       pass

   def destroy(self, request, *args, **kwargs):
       pass

So now my questions are:

1. If I'm not including def create(self, request) in the BookViewSet for example, I assume when I make a POST /book it will do nothing?
2. I'm struggling to understand what to put inside create , retrieve , update , destroy . For example if I make a PUT /book/123 with params: { isbn: '123', name: 'test'} , what do I need to do in def update to retrieve those data from request then update data to database?
3. I want to validate isbn and make sure it is unique/no duplicate before saving a record/update it to database, what should I do? I tried reading about Validator: https://www.django-rest-framework.org/api-guide/validators/ and looks like I need to update serializer:

class BookSerializer(serializers.ModelSerializer):
    name = serializers.CharField(source="BNAME")
    isbn = serializers.CharField(source="BISBN", validators=[<UniqueValidator(queryset=Book.objects.all())>])

        class Meta:
            model = models.Book
            fields = ('name', 'isbn')

Is this correct?

4. Similar to previous question, but now I want to make sure both isbn and name are unique together, I assume this is what I need to do:

class BookSerializer(serializers.ModelSerializer):
    name = serializers.CharField(source="BNAME")
    isbn = serializers.CharField(source="BISBN")

        class Meta:
            model = models.Book
            fields = ('name', 'isbn')

        validators = [
            UniqueTogetherValidator(
                queryset=Book.objects.all(),
                fields=('name', 'isbn')
            )
        ]

Is this correct?

Thanks

Answer 1

1. Your implementation of create(self, request, *args, **kwargs) will do nothing because the function is empty, it won't even return a response.

2. You don't have to put anything in create , list , retrieve , update , partial_update , and destroy because you're using the ModelViewSet class. The ModelViewSet provides a default implementation for all of them. If you really want to keep your overrides though you could use this implementation instead:

def create(self, request, *args, **kwargs):
    """You could add some documentation here"""
    return super().create(request, *args, **kwargs)

We're using super() because it is calling the parent implementation of create .

3. What you're doing is correct, let me just add that I would personally put a unique=True on the model field rather than adding the validator to the serializer. Rest Framework is smart enough that the serializer will inspect the model field and add the unique validation auto-magically. If you'd rather avoid the unique index on your database then your use of UniqueValidator is correct I think.

BISBN = models.CharField(db_column='BISBN', max_length=10, default=None, unique=True)

Keep in mind that you may be susceptible to race conditions without the unique index on the database. By this I mean that two books could be created/updated to have the same isbn if these operations occur at the same time.

4. Again, I would apply validation for multiple fields on the model using the Meta.unique_together attribute of the model, ie

class Book(models.Model):
    # fields...

    class Meta:
        unique_together = ("BNAME", "BISBN")

However if you don't want a unique index on your database for this then your use of the UniqueTogetherValidator is correct.

I have another tidbit I think is worth mentioning:

You view should have a lookup_url_kwarg and lookup_field to work with the isbn in your url definitions.

class BookViewSet(ModelViewSet):
    lookup_url_kwarg = "isbn"  # Use the 'isbn' from your url definition
    lookup_field = "BISBN"  # Use the `BISBN` field on your database/model to perform the lookups
    # ...

If you want to handle multiple url kwargs, then you will need to customize the get_queryset() method. In the following code I use isbn to filter the queryset, and then let the view lookup the book by name.

class BookViewSet(ModelViewSet):
    lookup_url_kwarg = "name"
    lookup_field = "BNAME"
    def get_queryset(self):
        # self.kwargs is a dictionary containing your url kwargs.
        isbn = self.kwargs.get("isbn")
        book_queryset = super().get_queryset()
        return book_queryset.filter(BISBN=isbn)