Django 按小时分组

Question

我在 Django 中有以下模型。

class StoreVideoEventSummary(models.Model):
    Customer = models.ForeignKey(GlobalCustomerDirectory, null=True, db_column='CustomerID', blank=True, db_index=True)
    Store = models.ForeignKey(Store, null=True, db_column='StoreID', blank=True, related_name="VideoEventSummary")
    Timestamp = models.DateTimeField(null=True, blank=True, db_index=True)
    PeopleCount = models.IntegerField(null=True, blank=True)

我想知道每小时进入商店的人数。

为了实现这一点，我尝试在 Timestamp 上按小时对行进行分组，并对 PeopleCount 列求和。

store_count_events = StoreVideoEventSummary.objects.filter(Timestamp__range=(start_time, end_time),
                                                       Customer__id=customer_id,
                                                       Store__StoreName=store)\
        .order_by("Timestamp")\
        .extra({
            "hour": "date_part(\'hour\', \"Timestamp\")"
        }).annotate(TotalPeople=Sum("PeopleCount"))

这似乎没有按小时对结果进行分组，它只是在查询集中的每一行中添加一个新列TotalPeople，其值与PeopleCount 相同。

Answer 1

把它分成两步

import itertools
from datetime import datetime


# ...

def date_hour(timestamp):
    return datetime.fromtimestamp(timestamp).strftime("%x %H")


objs = StoreVideoEventSummary.objects.filter(
    Timestamp__range=(start_time, end_time),
    Customer__id=customer_id,
    Store__StoreName=store
).order_by("Timestamp")

groups = itertools.groupby(objs, lambda x: date_hour(x.Timestamp))

# since groups is an iterator and not a list you have not yet traversed the list
for group, matches in groups:  # now you are traversing the list ...
    print(group, "TTL:", sum(1 for _ in matches))

这允许您按几个不同的标准进行分组

你只想要小时而不考虑日期，只需更改date_hour

def date_hour(timestamp):
   return datetime.fromtimestamp(timestamp).strftime("%H")

如果您想按星期几分组，您只需使用

def date_day_of_week(timestamp):
   return datetime.fromtimestamp(timestamp).strftime("%w %H")

并更新 itertools.groupby 的 lambda 以使用 date_day_of_week。

Answer 2

构建你的原始代码，你能试试吗：

store_count_events = StoreVideoEventSummary.objects.filter(Timestamp__range=(start_time, end_time), Customer__id=customer_id, Store__StoreName=store)\
    .extra({
        "hour": "date_part(\'hour\', \"Timestamp\")"
    })\
    .values("hour")\
    .group_by("hour")\
    .annotate(TotalPeople=Sum("PeopleCount"))

Answer 3

我知道我在这里迟到了，但从文档 https://docs.djangoproject.com/en/1.11/ref/models/querysets/#django.db.models.query.QuerySet.extra 那里得到提示

下面的过滤器应该适合你。

store_count_events = StoreVideoEventSummary.objects.filter(
    Timestamp__range=(start_time, end_time),
    Customer__id=customer_id,
    Store__StoreName=store
).order_by(
    'Timestamp'
).extra(
    select={
        'hour': 'hour(Timestamp)'
    }
).values(
    'hour'
).annotate(
    TotalPeople=Sum('PeopleCount')
)