Changed a test to be consistent when run in parallel.

This particular test was sometimes failing when running the test suite
in parallel. The `id` was different depending on the order the tests
were run. The test was incorrectly comparing model primary keys rather
than ensuring they didn't change.

tests/queries/tests.py

@@ -1645,6 +1645,8 @@ class Queries5Tests(TestCase):
     def test_ticket9848(self):
         # Make sure that updates which only filter on sub-tables don't
         # inadvertently update the wrong records (bug #9848).
+        author_start = Author.objects.get(name='a1')
+        ranking_start = Ranking.objects.get(author__name='a1')
 
         # Make sure that the IDs from different tables don't happen to match.
         self.assertQuerysetEqual(
@@ -1652,12 +1654,14 @@ class Queries5Tests(TestCase):
             ['<Ranking: 3: a1>']
         )
         self.assertEqual(
-            Ranking.objects.filter(author__name='a1').update(rank='4'),
+            Ranking.objects.filter(author__name='a1').update(rank=4636),
             1
         )
-        r = Ranking.objects.filter(author__name='a1')[0]
-        self.assertNotEqual(r.id, r.author.id)
-        self.assertEqual(r.rank, 4)
+
+        r = Ranking.objects.get(author__name='a1')
+        self.assertEqual(r.id, ranking_start.id)
+        self.assertEqual(r.author.id, author_start.id)
+        self.assertEqual(r.rank, 4636)
         r.rank = 3
         r.save()
         self.assertQuerysetEqual(