CODEWITHSUNDEEP
djangodjango-modelsdjango-aggregationdjango-annotate
Niladry Kar
Niladry Kar

Reputation: 1203

Django: Problem in doing complex annotation and aggregation

This is model:

class Purchase(models.Model):
    date           = models.DateField(default=datetime.date.today,blank=False, null=True)
    total_purchase = models.DecimalField(max_digits=10,decimal_places=2,blank=True, null=True)

I want to perform a month wise calculation of "total_purchase" within a specific daterange in such a way that if there is no purchase in a month the total purchase should be the previous month purchase value And if there is purchase in two months then total purchase will the addition of those two...

Example:

Suppose the date range given by user is from month of April to November.

If there is a Purchase of $2800 in month of April and $5000 in month of August and $6000 in month of October.

Then the output will be like this:

April      2800
May        2800
June       2800
July       2800
August     7800  #(2800 + 5000)
September  7800
October    13800 #(7800 + 6000)
November   13800

Any idea how to perform this in django queries?

Thank you

According to the answer given by Mr.Raydel Miranda. I have done the following

import calendar
import collections
import dateutil

start_date = datetime.date(2018, 4, 1)
end_date = datetime.date(2019, 3, 31)

results = collections.OrderedDict()

result = Purchase.objects.filter(date__gte=start_date, date__lt=end_date).annotate(real_total = Case(When(Total_Purchase__isnull=True, then=0),default=F('tal_Purchase')))

date_cursor = start_date

while date_cursor < end_date:
    month_partial_total = result.filter(date__month=date_cursor.month).agggate(partial_total=Sum('real_total'))['partial_total']

    results[date_cursor.month] = month_partial_total

    if month_partial_total == None:
            month_partial_total = int(0)
    else:
            month_partial_total = month_partial_total

    date_cursor += dateutil.relativedelta.relativedelta(months=1)

    return results

But now the output is coming like this(from the example above):

April      2800
May        0
June       0
July       0
August     5000
September  0
October    6000
November   0

Do anyone have any idea how to add between the months... I want to do something like

e = month_partial_total + month_partial_total.next

I want to add the next iteration value of every month_partial_total. I think this will solve my problem..

Any idea anyone how to perform this in django?

Thank you

Upvotes: 4

Views: 398

Answers (2)

Niladry Kar
Niladry Kar

Reputation: 1203

Solution

According to the answer given by Mr.Raydel Miranda, Finally I got the solution to my problem...

I have done the following in my views and it worked very nicely:

import datetime
import calendar
import collections
import dateutil
start_date = datetime.date(2018, 4, 1)
end_date = datetime.date(2019, 3, 31)
results = collections.OrderedDict()
result = Purchase.objects.filter(date__gte=start_date, date__lt=end_date).annotate(real_total = Case(When(Total_Purchase__isnull=True, then=0),default=F('Total_Purchase')))
date_cursor = start_date
z = 0
while date_cursor < end_date:
    month_partial_total = result.filter(date__month=date_cursor.month).aggregate(partial_total=Sum('real_total'))['partial_total']
    # results[date_cursor.month] = month_partial_total
    if month_partial_total == None:
        month_partial_total = int(0)
        e = month_partial_total
    else:
        e = month_partial_total

    z = z + e

    results[date_cursor.month] = z

    date_cursor += dateutil.relativedelta.relativedelta(months=1)

 return results

Thank you everyone.

Upvotes: 0

Raydel Miranda
Raydel Miranda

Reputation: 14360

I've noted two things in your question:

  1. Results are ordered by month.
  2. Total purchase can be blank or null.

Based on those things I'll propose this approach:

You could get the total for a given month, you just need to handle the case where the total_pushase is null (as a side note, it does not make any sense to have an instance of Purchase where total_purchase is null, at least it must be 0).

Read about Django Conditional expressions to learn more about When and Case.

# Annotate the filtered objects with the correct value (null) is equivalent
# to 0 for this requirement.

result = Purchase.objects.filter(date__gte=start_date, date__lt=end_date).annotate(
    real_total = Case(
        When(total_purchase__isnull=True, then=0),
        default=F('total_purchase')
    )
)

# Then if you want to know the total for a specific month, use Sum.
month_partial_total = result.filter(
    date__month=selected_month
).aggregate(
    partial_total=Sum('real_total')
)['partial_total']

You could use this in a function to achieve you wanted result:

import calendar
import collections
import dateutil

def totals(start_date, end_date):
    """
    start_date and end_date are datetime.date objects.
    """

    results = collections.OrderedDict()  # Remember order things are added.

    result = Purchase.objects.filter(date__gte=start_date, date__lt=end_date).annotate(
        real_total = Case(
            When(total_purchase__isnull=True, then=0),
            default=F('total_purchase')
        )
    )

    date_cursor = start_date
    month_partial_total = 0
    while date_cursor < end_date:
        # The while statement implicitly orders results (it goes from start to end).
        month_partial_total += result.filter(date__month=date_cursor.month).aggregate(
            partial_total=Sum('real_total')
        )['partial_total']


        results[date_cursor.month] = month_partial_total

        # Uncomment following line if you want result contains the month names
        # instead the month's integer values.
        # result[calendar.month_name[month_number]] = month_partial_total

        date_cursor += dateutil.relativedelta.relativedelta(months=1)

    return results

Since Django 1.11 might be able to solve this problem SubQueries, but I've never used it for subquery on the same model.

Upvotes: 7

Related Questions