Default QuerySet for DjangoObjectType has no effect with graphene.List

The documentation describes the use of a default QuerySet on ObjectType level.

The unit test for this functionality makes use of DjangoConnectionField.

Using graphene.List instead ignores the existence of the default QuerySet.

class CurrencyType(DjangoObjectType):
    class Meta:
        model = Currency

    @classmethod
    def get_queryset(cls, queryset, info):
        print(cls)
        print(queryset)
        return queryset.filter(symbol__isnull=False)

all_currencies = graphene.List(CurrencyType)

def resolve_all_currencies(self, info):
    return Currency.objects.all()

The get_queryset method is never called, and the filter is not applied.

Please refer to #698 for some history on this.

[jkimbo]

Something that still doesn’t work as I hoped is the following: When Currency from my example is a related field in another table, let’s say Fund, a query for funds does not call get_queryset in the CurrencyType class.

This issue should actually be fixed as part of my PR. I’ll add a test to confirm.

[pors]

Awesome @jkimbo! I have it working now for the example I provided (I don’t need a resolver on the query level here). I think it actually makes sense as it is called the “Default QuerySet”.

Something that still doesn’t work as I hoped is the following: When Currency from my example is a related field in another table, let’s say Fund, a query for funds does not call get_queryset in the CurrencyType class.

If that would be possible it provides an easy way to protect from returning too much data for specific tables (this is especially important for the User table being a related field).