Django `bulk_update()` : Update Different Fields for Each Record in a Single Query

You can do this, in fact the Django ORM will, for .bulk_update(..) essentially make a query that looks like:

from django.db.models import Case, Value, When

model.objects.filter(pk__in=[rec1.pk, rec2.pk]).update(
    name=Case(
        When(pk=rec1.pk, then=Value(rec1.name)),
        When(pk=rec2.pk, then=Value(rec2.name)),
        default=F('name'),
    ),
    age=Case(
        When(pk=rec1.pk, then=Value(rec1.age)),
        When(pk=rec2.pk, then=Value(rec2.age)),
        default=F('age'),
    ),
)

This is not very efficient: most databases will fall back to a "linear" search through the CASE-WHENs, and thus updating the records might be very slow.

so we can limit the cases to only the ones we want to update with:

model.objects.filter(pk__in=[rec1.pk, rec2.pk]).update(
    name=Case(When(pk=rec1.pk, then=Value(rec1.name)), default=F('name')),
    age=Case(When(pk=rec2.pk, then=Value(rec2.age)), default=F('age')),
)

TLDR;

We want to update different fields for different records in a single bulk_update() operation. This is indeed a limitation with Django's current implementation.

The key issue is that Django's bulk_update() method requires you to specify a fixed list of fields that will be applied to ALL objects in your update batch. There's no built-in way to say "only update the 'name' field for record 1, but only update the 'age' field for record 2" in a single operation.

When we use something like

rec1 = MyModel(id=1, name="John")  # Only 'name' is changed
rec2 = MyModel(id=2, age=30)       # Only 'age' is changed
model.objects.bulk_update([rec1, rec2], fields=['name', 'age'])

Django will generate SQL that updates both fields for both records, which isn't what you want.

It looks something like

UPDATE my_table SET 
  name = CASE WHEN id=1 THEN 'John' WHEN id=2 THEN name END,
  age = CASE WHEN id=1 THEN age WHEN id=2 THEN 30 END
WHERE id IN (1,2)

It's updating both fields for both records, even though you only wanted to update one field per record. This is the inherent limitation.

As of now, there are several approaches to handle this scenario:

Group objects by the fields that need updating and perform separate bulk_update() calls:

name_updates = [obj for obj in all_objects if obj.name_changed]
age_updates = [obj for obj in all_objects if obj.age_changed]

MyModel.objects.bulk_update(name_updates, ['name'])
MyModel.objects.bulk_update(age_updates, ['age'])

Unfortunately, Django just doesn't have built-in support for this use case. According to multiple discussions, including those on Reddit, "Django and most ORMs don't bother with this. They either use all fields in the query or just the ones you specify"

There have been one discussion Reduce the "Case-When" sequence for a bulk_update when the values for a certain field are the same. an year ago about the same. You are free to use this approach with a custom manager/queryset.

I think the most efficient way to update the objects is to use bulk_update twice. Since the sets in your example are disjoint, each object will be updated only once.

Running a single bulk_update would require an if condition to handle different cases, which introduces additional processing overhead

.