Question

I’m working with a model:

class Foo(models.Model):
    name = models.CharField(max_length=255)
    bars = models.ManyToManyField('Bar')

In a view I have access to a list of Barobjects and need to get all Foo objects that have any of the Bar objects in their list of bars, so I do this:

foos = Foo.objects.filter(bars__in=list_of_bars)

The issue is that there are duplicates, if a Foo has 2 bars, and both of those bars are in my list_of_bars, which a simple distinct solves:

foos = Foo.objects.distinct().filter(bars__in=list_of_bars)

That’s all good and well, except that adding DISTINCT to the query makes it very slow, due to there being 2 million Foo objects in the database.

With all that said, what way(s) can you think of that don’t use DISTINCT, but achieve the same result set? If it involves altering models, that’s OK.

Answer 1

You could always select uniques in python:

foos = set(Foo.objects.filter(bars__in=list_of_bars))

You need query set. So, maybe (just maybe) this quick nasty hack could help You. It is ugly, but maybe it is fast (not sure about that).

def get_query_set(self):
    foos = set(Foo.objects.filter(bars__in=list_of_bars))
    return Foo.objects.filter(id__in=[f.id for f in foos])