Get a set of objects using an intermediate model

I want to get all Geoms associated with a specific content_object (see the function I'm trying to plot at the bottom get_geoms_for_obj ()

class Geom(models.Model):
    ...

class GeomRelation(models.Model):
    ''' For tagging many objects to a Geom object and vice-versa'''

    geom = models.ForeignKey(Geom)
    content_type = models.ForeignKey(ContentType)
    object_id = models.PositiveIntegerField()
    content_object = generic.GenericForeignKey()

def get_geoms_for_object(obj):
    ''' takes an object and gets the geoms that are related

    '''
    ct = ContentType.objects.get_for_model(obj)
    id = obj.id
    grs = GeomRelation.objects.filter( content_type=ct, object_id=id )
    # how with django orm magic can I build the queryset instead of list
    # like below to get all of the Geom objects for a given content_object
    geoms = []
    for gr in grs:
        geoms.append(gr.geom)
    return set(geoms)
    # A set makes it so that I have no redundant entries but I want the
    # queryset ordering too .. need to make it a queryset for so many reasons...

      

+2


source to share


3 answers


Duh,



return Geom.objects.filter(geomrelation__in=grs)

      

+1


source


If your Geom class has a generic.GenericRelation () property, you can get objects through the standard inverse relationship.

class Geom(models.Model):
    ...
    geom_relations = generic.GenericRelation(GeomRelation)

      



This is a Python-only property and does not require database changes. Now, to get geom_relations, you simply do:

geom.geom_relations.all()

      

+1


source


If you pass one QuerySet object as a filter argument to another, the ORM will use nested select statements. It is best to join table relationships in search movies; then the ORM will use a simple WHERE clause in the JOIN, which it should do anyway. The performance difference is significant:

In [2]: from django.db import connection
In [3]: from app.models import *
In [4]: ct = ContentType.objects.get_for_model(Thing)
In [5]: grs = GeomRelation.objects.filter( content_type=ct, object_id=2 )
In [6]: 
In [7]: # slow method
In [8]: list(Geom.objects.filter(geomrelation__in=grs));
In [9]: connection.queries[-1]
Out[9]: 
{'sql': u'SELECT "app_geom"."id" FROM "app_geom" INNER JOIN "app_geomrelation" ON ("app_geom"."id" = "app_geomrelation"."geom_id") WHERE "app_geomrelation"."id" IN (SELECT U0."id" FROM "app_geomrelation" U0 WHERE (U0."content_type_id" = 10  AND U0."object_id" = 2 ))',
 'time': '0.140'}
In [10]: 
In [11]: # fast method
In [12]: list(Geom.objects.filter(geomrelation__content_type=ct,
   ....:                         geomrelation__object_id=2));
In [13]: connection.queries[-1]
Out[13]: 
{'sql': u'SELECT "app_geom"."id" FROM "app_geom" INNER JOIN "app_geomrelation" ON ("app_geom"."id" = "app_geomrelation"."geom_id") WHERE ("app_geomrelation"."object_id" = 2  AND "app_geomrelation"."content_type_id" = 10 )',
 'time': '0.001'}

      

0


source







All Articles