Let's imagine a model called Roll
. It stores the outcome of rolling a six-sided die (D6):
class Roll(models.Model):
outcome = models.PositiveSmallIntegerField('Roll', null=False, blank=False, default=1)
There are many rolls, for example:
print(list(Roll.objects.all().values_list('outcome', flat=True)))
>>> [1, 5, 6, 3, 5, 4, 4, 3, 2]
Now, how should I go about to get the latest N rows whose Sum('outcome') reaches an arbitrary amount, without looping in some way or another, until that amount is reached?
If we imagine the rolls:
pk | outcome | (accumulated sum)
1 | 3 | 3
2 | 2 | 5
3 | 6 | 11
4 | 1 | 12
5 | 5 | 17
6 | 4 | 21
7 | 3 | 24
8 | 4 | 29
9 | 5 | 34
10 | 1 | 35
and the arbitrary amount 20
, then the query should pick the pk 6
as the accumulated sum has now reached the amount needed.
Could something similar to the below work?:
amount = 100
Roll.objects.annotate(
accumulated_sum=Subquery(
Roll.objects.filter(
pk__lte=OuterRef('pk')
).values('pk').order_by('pk').annotate(
sum=Sum('outcome', distinct=True)
).values(
'sum'
)[:1]
)
).filter(
accumulated_sum__gte=amount
)
CodePudding user response:
This might help annotating accumuated_sum:
from django.db.models import Subquery, IntegerField
class SQSum(Subquery):
output_field = IntegerField()
template = "(SELECT sum(outcome) from (%(subquery)s) _sum)"
accumulated_subquery = SQSum(
Roll.objects.filter(
pk__lte=OuterRef('pk')
).values("outcome")
)
Roll.objects.annotate(accumulated_sum=accumulated_subquery).filter(accumulated_sum__gte=amount)