Django ORM:获取具有相应其他字段值的字段的最大值

Django ORM: Get maximum value of a field with corresponding other fields values

我有这个 Table(计数器):

cell_id tftralacc tfnscan thtralacc thnscan date_time
13997 10 360 94 360 2022-02-22 00:00:00+01
13997 0 360 0 360 2022-02-22 01:00:00+01
13997 0 360 0 360 2022-02-22 02:00:00+01
13997 0 360 0 360 2022-02-22 03:00:00+01
13997 36 360 83 360 2022-02-22 04:00:00+01
13997 0 360 2 360 2022-02-22 05:00:00+01
13997 1 360 15 360 2022-02-22 06:00:00+01
13997 11 360 159 360 2022-02-22 07:00:00+01
13997 21 360 409 360 2022-02-22 08:00:00+01
13997 25 360 1282 360 2022-02-22 09:00:00+01
13997 20 360 1201 360 2022-02-22 10:00:00+01
13997 30 360 1381 360 2022-02-22 11:00:00+01
13997 42 360 924 360 2022-02-22 12:00:00+01
14000 1 360 36 360 2022-02-22 00:00:00+01
14000 0 360 0 360 2022-02-22 01:00:00+01
14000 1 360 0 360 2022-02-22 02:00:00+01
14000 0 360 2 360 2022-02-22 03:00:00+01
14000 0 360 0 360 2022-02-22 04:00:00+01
14000 0 360 12 360 2022-02-22 05:00:00+01
14000 3 360 4 360 2022-02-22 06:00:00+01
14000 24 360 123 360 2022-02-22 07:00:00+01
14000 31 360 374 360 2022-02-22 08:00:00+01
14000 18 360 620 360 2022-02-22 09:00:00+01
14000 38 360 1616 360 2022-02-22 10:00:00+01
14000 36 360 1410 360 2022-02-22 11:00:00+01
14000 24 360 957 360 2022-02-22 12:00:00+01

我想获取最大流量的具体date_time值(这是根据字段计算的tftralacctfnscanthtralaccthnscan)对于每个cell_id.

我已经设法通过使用 Django 的 QuerySet 的 annotate()group_by() 函数为每个 cell_id 获得最大值API:

result = Counters.objects.filter(
    date_time__gte = date_start,
    date_time__lte = date_end
).annotate(
    # calculate the traffic for each row.
    traffic = Case(
        When(Q(tfnscan=0) or Q(thnscan=0), then=0),
        default = Round((F('tftralacc')*1.0/F('tfnscan')) + 
                        (F('thtralacc')*1.0/F('thnscan')), 2),
        output_field=FloatField()
    )
).order_by('cell_id').values(
    # Group by cell_id.
    'cell_id'
).order_by().annotate(
    # calculate the max traffic for the grouped Cells.
    max_traffic = Max('traffic')
)

为每个 date_time 计算的 流量 如下所示:

我的代码成功了returns每个最大流量cell_id:

cell_id max_traffic
13997 3.92
14000 4.59

但我的目标是为每个最大值获取相应的 date_time 值。像这样:

cell_id max_traffic date_time
13997 3.92 2022-02-22 11:00:00+01
14000 4.59 2022-02-22 10:00:00+01

cell_id date_time
13997 2022-02-22 11:00:00+01
14000 2022-02-22 10:00:00+01

因为最大值只是获得 date_time 的手段,而不是目标。

注意:有一个问题描述了我的问题,但它的答案是指一个变通解决方案,这对我的问题是不可能的。

使用 models.Subquery with models.OuterRef 加入 cell_id 字段。然后用queryset.annotate()注解子查询max_traffic。最后,使用 queryset.filter() 到 select 具有 traffic 等于 max_traffic 的行,并使用 .distinct() 删除重复行。

counters_with_traffic = Counters.objects.filter(
    date_time__gte=date_start,
    date_time__lte=date_end
).annotate(
    # calculate the traffic for each row.
    traffic=Case(
        When(Q(tfnscan=0) | Q(thnscan=0), then=0),
        default=Round((F('tftralacc') * 1.0 / F('tfnscan')) +
                      (F('thtralacc') * 1.0 / F('thnscan')), 2),
        output_field=models.FloatField()
    )
)

counters_with_max_traffic = counters_with_traffic.order_by('cell_id').values(
    # Group by cell_id.
    'cell_id'
).order_by().annotate(
    # calculate the max traffic for the grouped Cells.
    max_traffic=Max('traffic'),
).filter(cell_id=models.OuterRef("cell_id")).values("max_traffic")

result = counters_with_traffic.annotate(
    max_traffic=models.Subquery(counters_with_max_traffic),
).filter(
    traffic=models.F("max_traffic")
).values(
    "cell_id", "max_traffic", "date_time"
).distinct("cell_id", "max_traffic")