[Question] Best practice for nested data in POST

Hi,

Imagine I have one team with a list of users. So when I get the data, it will be something like this: {id: 1, name:'team', users:[{id:1, username='john'}]} the schema would be simple:

class BaseUserSchema(ModelSchema):
    class Meta:
        model = User
        fields = (
            User.id.key,
            User.username.key
        )
       dump_only = (
            User.id.key,
        )

class BaseTeamSchema(ModelSchema):
    users = base_fields.List(base_fields.Nested(BaseUserSchema))
    class Meta:
        model = Team
        fields = (
            Team.id.key,
            Team.name.key,
            Team.users.key,
        )
       dump_only = (
            Team.id.key,
        )

Now my question is, when I want to add a team, with some users, I want to send something like this: {name:'team B', users:[{id:1}, {id:2}, {id:3}]} with already existing users. But, it will not have the users id, since it is in dump_only fields. Even through I override it in CreateTeamParameters with users = base_fields.List(base_fields.Nested(BaseUserSchema(only=['id']))) it doesn’t override the dump_only.

What would be the best practice in order to Post nested schemas?

Thank you!

[0xecute]

Okay, we were almost right. Using dump_only=[] should work, but the following code in marshmallow: self.dump_only = set(dump_only) or set(self.opts.dump_only) kind of fail. Since set([]) is equal to 0, it will go with the or. So you should just not use an empty array and do something like this for example:

users = base_fields.List(base_fields.Nested(BaseUserSchema(only=['id'], dump_only=[None]))) Not the perfect solution, but it works and keep your basic schema secured by not accepting ID rewrite.

Thank you again for the project.

[frol]

I am sorry, I won’t be able to help you with this.