Why does a Django QuerySet with a Q () expression return duplicate values?

Question

I have the following two Django models:

from mptt.models import MPTTModel, TreeForeignKey
from django.db import models
from django.db.models import Q

class Model1(MPTTModel):
    random_field = models.IntegerField()
    parent = TreeForeignKey('self', null=True, blank=True)


class Model2(models.Model):
    model_1 = models.ManyToManyField(Model1)

    @staticmethod
    def descendants_queryset(model1):
        q = Q()
        for curr_descendant in model1.get_descendants:
            q |= Q(model_1=curr_descendant)
        return q

I created such instances:

>>> a = Model2.objects.create()
>>> b = Model1.objects.create(random_field=1, parent=None)
>>> c = Model1.objects.create(random_field=2, parent=b)
>>> d = Model1.objects.create(random_field=3, parent=b)
>>> a.model_1.add(c)
>>> a.pk
3

When I do a regular query filter and when I use the Q () expression, it gives the same results (as expected):

>>> [i.pk for i in Model2.objects.filter(pk=3)]
[3]
>>> [i.pk for i in Model2.objects.filter(Model2.descendants_queryset(b), pk=3)]
[3]

But when I add another instance of Model1 to the ManyToMany relation, I see weird duplication only when I filter using the Q () expression:

>>> a.model_1.add(d)
>>> [i.pk for i in Model2.objects.filter(pk=3)]
[3]
>>> [i.pk for i in Model2.objects.filter(Model2.descendants_queryset(b), pk=3)]
[3, 3]

I am confused why this duplication occurs. It seems to me that this is a mistake. I will obviously get around it by adding .distinct()to the request. But it does not seem to be necessary. Why is this happening and what is the appropriate solution?

Answer 1

, a, , :

>>> 4 = Model1.objects.create(random_field=3, parent=b)
>>> a.model_1.add(e)
>>> [i.pk for i in Model2.objects.filter(Model2.descendants_queryset(b), pk=3)]
[3, 3, 3]

, ...

, , Q()-query descendants_queryset() ORed, , b , ( multple Model1).

SQL Model2.objects.filter(Model2.descendants_queryset(b)), :

>>> Model2.objects.filter(Model2.descendants_queryset(b)).query.sql_with_params()
(u'SELECT "Foo_model2"."id" FROM "Foo_model2" LEFT OUTER JOIN "Foo_model2_model_1" ON ("Foo_model2"."id" = "Foo_model2_model_1"."model2_id") WHERE ("Foo_model2_model_1"."model1_id" = %s OR "Foo_model2_model_1"."model1_id" = %s OR "Foo_model2_model_1"."model1_id" = %s)', (17, 18, 19))

:

SELECT "Foo_model2"."id"
FROM "Foo_model2"
  LEFT OUTER JOIN "Foo_model2_model_1"
  ON ("Foo_model2"."id" = "Foo_model2_model_1"."model2_id")
WHERE ("Foo_model2_model_1"."model1_id" = 17
  OR "Foo_model2_model_1"."model1_id" = 18
  OR "Foo_model2_model_1"."model1_id" = 19)

, , q |= Q(model_1=curr_descendant) OR, , ( Model2, ManyToMany Model1), - . .

pk=3, , PK (3).

Model2 c model1 ManyToMany-reference, :

>>> a2 = Model2.objects.create()
>>> a2.model_1.add(c)
>>> [i.pk for i in Model2.objects.filter(Model2.descendants_queryset(b))]
[3, 3, 3, 4]

Model2 , model1.

, .distinct() .