I have a model:
class Zone(models.Model):
name = models.CharField(max_length=128)
users = models.ManyToManyField(User, related_name=\'zones\', null=T
Just restating what Tomasz said.
There are many examples of FOO__in=...
style filters in the many-to-many and many-to-one tests. Here is syntax for your specific problem:
users_in_1zone = User.objects.filter(zones__id=<id1>)
# same thing but using in
users_in_1zone = User.objects.filter(zones__in=[<id1>])
# filtering on a few zones, by id
users_in_zones = User.objects.filter(zones__in=[<id1>, <id2>, <id3>])
# and by zone object (object gets converted to pk under the covers)
users_in_zones = User.objects.filter(zones__in=[zone1, zone2, zone3])
The double underscore (__) syntax is used all over the place when working with querysets.
Note that if the user may be in multiple zones used in the query, you may probably want to add .distinct()
. Otherwise you get one user multiple times:
users_in_zones = User.objects.filter(zones__in=[zone1, zone2, zone3]).distinct()
another way to do this is by going through the intermediate table. I'd express this within the Django ORM like this:
UserZone = User.zones.through
# for a single zone
users_in_zone = User.objects.filter(
id__in=UserZone.objects.filter(zone=zone1).values('user'))
# for multiple zones
users_in_zones = User.objects.filter(
id__in=UserZone.objects.filter(zone__in=[zone1, zone2, zone3]).values('user'))
it would be nice if it didn't need the .values('user')
specified, but Django (version 3.0.7) seems to need it.
the above code will end up generating SQL that looks something like:
SELECT * FROM users WHERE id IN (SELECT user_id FROM userzones WHERE zone_id IN (1,2,3))
which is nice because it doesn't have any intermediate joins that could cause duplicate users to be returned