django annotate weird behavavior (group by model.id)

Question

在我的DRF API中，我有一个这样的视图

class ActivityAPI(viewsets.ModelViewSet):
    authentication_classes = (SessionAuthentication, TokenAuthentication)
    serializer_class = ActivitySerializer
    queryset = Activity.objects.order_by('-id').all()
    filter_backends = (DjangoFilterBackend,)
    filter_class = ActivityFilter
    filter_fields = ('name', 'ack', 'developer', 'ack_date', 'ack_by', 'verb')    
    def get_count(self, request, *args, **kwargs):
        queryset = self.filter_queryset(self.get_queryset())
        if CASE_1:
            queryset = queryset.values('verb').annotate(count=Count('verb'))
        if CASE_2:
            queryset = Activity.objects.values('verb').annotate(count=Count('verb'))
        return Response(data=queryset)

在CASE_2，我得到了我所期望的，相当于SQL查询

SELECT `activity_activity`.`verb`, COUNT(`activity_activity`.`verb`) AS `count` FROM `activity_activity` GROUP BY `activity_activity`.`verb` ORDER BY NULL

但是当涉及到CASE_1时，注释功能按activity.id对查询集进行分组，即

SELECT `activity_activity`.`verb`, COUNT(`activity_activity`.`verb`) AS `count` FROM `activity_activity` GROUP BY `activity_activity`.`id` ORDER BY `activity_activity`.`id` DESC`

注意 我需要基于 url 的 API 和聚合过滤数据

Answer 1

组之所以如此之小，是因为您在原始查询集上具有.order_by('-id')。您可以通过空order_by或仅按分组中使用的字段排序来修复它：

if CASE_1:
    queryset = queryset.values('verb').annotate(count=Count('verb')).order_by('verb')

阅读有关聚合 - 与默认排序或 order_by（） 的交互的信息。