Django通用外键过滤(v1.5和v1.6之间的区别)

Question

我有以下概念模型:

class GenericAbstractBase(models.Model):
    name = models.CharField(max_length=255)
    staff = generic.GenericRelation(
        "Staff",
        content_type_field="content_type",
        object_id_field="object_id",
    )

    class Meta:
        abstract = True


class GenericModelA(GenericAbstractBase):
    extra_a = models.CharField(max_length=255)


class GenericModelB(GenericAbstractBase):
    extra_b = models.CharField(max_lenth=10)


class Staff(models.Model):

    first_name = models.CharField(max_length=255)
    last_name = models.CharField(max_length=255)
    active = models.CharField(max_length=10, choices = ACTIVE_CHOICES)

    limit = models.Q(app_label='staff', model='genericmodela') | models.Q(app_label='staff', model='genericmodelb')
    content_type = models.ForeignKey(ContentType, limit_choices_to=limit)
    object_id = models.PositiveIntegerField()
    generic_object = generic.GenericForeignKey("content_type", "object_id")

在Django v1.4和Django v1.5中,以下查询工作正常:

>>> ctype = ContentType.objects.get_for_model(GenericModelA)
>>> Staff.objects.filter(
        content_type=ctype,
        genericmodela__name__icontains="a"
    )
>>> [<Staff: Abbott, Kaylee>, <Staff: Adams, Kay>, ... ]

它产生的SQL(sqlite)看起来像:

SELECT 
    "staff_staff"."id", "staff_staff"."first_name","staff_staff"."last_name",
"staff_staff"."active","staff_staff"."content_type_id" ,"staff_staff"."object_id"
FROM "staff_staff"
INNER JOIN "staff_staff" T2 ON ("staff_staff"."id" = T2."id")
INNER JOIN "staff_genericmodela" ON (T2."object_id" = "staff_genericmodela"."id")
WHERE (
"staff_genericmodela"."name" LIKE % a % ESCAPE \ '\\\'
AND "staff_staff"."content_type_id" = 11
)

但是在Django 1.6中,查询失败并出现FieldError:

FieldError: Cannot resolve keyword 'genericmodela' into field. Choices are: active, content_type, department, first_name, id, last_name, object_id, position

发行说明中的以下声明可能是相关的:

Django 1.6包含对ORM的许多更改.这些变化大多分为三类:

1. 错误修复(例如,通用关系的正确连接子句,查询组合,连接提升和连接修剪修复)

我的问题是,在Django 1.6中发生了什么变化导致了这种情况的破坏？我是否坚持extra在Python中使用或执行此类过滤？

Answer 1

我在这里发现了一些有趣的信息。

作为解决方法，您可以执行以下操作：

ctype = ContentType.objects.get_for_model(GenericModelA)

pk_list = Staff.objects.filter(
    content_type=ctype
).values_list('object_id', flat=True)

GenericModelA.objects.filter(pk__in=pk_list, name__icontains="a")