My attempt to use prefetch related to access M2M data on a massive table with a massive related table

rlybig_prefetch_related.py

from myapp.models import BigAssModel

total_rows = BigAssModel.objects.count()
batch_size = 5000
cut_points = [i*batch_size for i in range(1,total_rows/batch_size)]
cut_points.insert(0, 0)
cut_points.append(total_rows)
for i, c in enumerate(cut_points):
    if c == cut_points[-1]:
        print "DONE"
    else:
        print "%s-%s" % (c, cut_points[i+1])
        qs = BigAssModel.objects.all().prefetch_related("relatedobj_set")[c:cut_points[i+1]]
        for obj in qs:
          # Do all kinds of stuff with relatedobj_set, but just to have a dumb example
          obj.relatedobj_count = len(obj.relatedobj_set.all())
          obj.save()