CODEWITHSUNDEEP
pythondjango
Mr.D
Mr.D

Reputation: 7903

Django, how to group models by date?

Lets say that I have MyModel that has created_at and name fields. created_at is DateTime.

Lets say that I have the following model objects:

<id: 1, name: A, created_at: 04.06.2020T17:49>
<id: 2, name: B, created_at: 04.06.2020T18:49>
<id: 3, name: C, created_at: 05.06.2020T20:00>
<id: 4, name: D, created_at: 06.06.2020T19:20>
<id: 5, name: E, created_at: 06.06.2020T13:29>
<id: 6, name: F, created_at: 06.06.2020T12:55>

I want to make query that will return to me these models in this order:

[
    04.06.2020: [<id: 1, name: A, created_at: 04.06.2020T17:49>, <id: 2, name: B, created_at: 04.06.2020T18:49>],
    05.06.2020: [<id: 3, name: C, created_at: 05.06.2020T20:00>]
    06.06.2020: [<id: 4, name: D, created_at: 06.06.2020T19:20>, <id: 5, name: E, created_at: 06.06.2020T13:29>, <id: 6, name: F, created_at: 06.06.2020T12:55>]
]

I want to group all models by created_at field, but only using Date part of DateTime field. I know that I can make that kind of result just by using python loops, but are there any Django ORM way to solve this problem?

Upvotes: 6

Views: 1025

Answers (4)

Simon Charette
Simon Charette

Reputation: 5116

The ORM itself cannot return list of models but you can use itertools.groupby for this purpose assuming you order the dataset at the database level first.

from itertools import groupby
from operator import attrgetter

from django.db.models.functions import TruncDate

queryset = MyModel.objects.annotate(
    created_at_date=TruncDate('created_at'),
).order_by('created_at')
groupedset = groupby(queryset, attrgetter('created_at_date'))

You'll then be able able to iterate over groupedset using

for date, objects in groupedset:
    ...

Upvotes: 5

ACimander
ACimander

Reputation: 1979

I'm afraid this is currently not completely possible, at least to my knowledge (without some really deep modifications within the Django ORM).

What you can do, however, is doing date conversions and aggregations within the database:

This example assumes your database supports the to_char function.

qs = MyModel.objects.all()
qs.annotate(
  created_at_date=models.Func(
    models.F('created_at'),
    models.Value("'dd.MM.yyyy'"),
    function='to_char',
  )
).values('created_at_date').aggregate(
    # do required aggregations like Count or Sum
)

[edit] some fixes + as bryan60 correctly pointed out, you can also use Trunc:

from django.db.models.functions import Trunc
MyModel.objects.all().annotate(
    created_at_trunc=Trunc('created_at', 'day', output_field=models.DateTimeField())
).values('created_at_trunc')

Upvotes: 2

hardik24
hardik24

Reputation: 1058

You can try following code:

samples = MyModel.objects.all()
result = {}
for sample in samples:
    date_string = sample.created_at.strftime("%m.%d.%Y")
    if date_string in result:
        result[date_string].append(sample)
    else:
        result[date_string] = [sample]

It will return output like following:

{
'06.07.2020': [<Sample: MyModel object (1)>],
'05.07.2020': [<Sample: MyModel object (2)>, <Sample: MyModel object (3)>]
}

Basically, what it does it it first fetches all the rows and then group it at python level into a dictionary.

Upvotes: 2

rahimcodes
rahimcodes

Reputation: 1

Not quite understanding but if you have the database created under models.py then you can create a meta class that shows ordering 

Class myModel(models.Model):
    #fields...

     Class Meta:
         ordering = ('-created_at')

Upvotes: -2

Related Questions