#34907: loaddata crashes on objects with natural keys when don't exist on passed
database.
-------------------------------------+-------------------------------------
Reporter: Florian | Owner: nobody
Type: Bug | Status: new
Component: Core | Version: 4.2
(Serialization) |
Severity: Normal | Resolution:
Keywords: natural key, multi- | Triage Stage: Accepted
db, loaddata |
Has patch: 0 | Needs documentation: 0
Needs tests: 0 | Patch needs improvement: 0
Easy pickings: 0 | UI/UX: 0
-------------------------------------+-------------------------------------
Comment (by Florian):
Thanks for looking into it. The patch seems to fix it for me (filling in
some small details like importing the router):
{{{#!diff
diff --git a/django/core/serializers/python.py
b/django/core/serializers/python.py
index 7ec894aa00..2ef5a1765a 100644
--- a/django/core/serializers/python.py
+++ b/django/core/serializers/python.py
@@ -6,7 +6,7 @@ other serializers.
from django.apps import apps
from django.core.serializers import base
-from django.db import DEFAULT_DB_ALIAS, models
+from django.db import DEFAULT_DB_ALIAS, models, router
from django.utils.encoding import is_protected_type
@@ -118,6 +118,8 @@ def Deserializer(
else:
raise
data = {}
+ if not router.allow_migrate(using, Model):
+ continue
if "pk" in d:
try:
data[Model._meta.pk.attname] =
Model._meta.pk.to_python(d.get("pk"))
}}}
The test script outputs 'Ok' when running all tests with this patch
applied.
I can't really figure out what would be a logical unit test for the bug.
I've been looking at `tests/serializers/test_natural.py` for inspiration.
But I guess I'm not familiar enough with this part to come up with a good
test case that fails without this patch. What I've tried is modifying
`natural_key_test` to use a different database (modifications marked with
comments):
{{{
class TestRouter:
def allow_migrate(self, db, app_label, model_name=None, **hints):
return False
@override_settings(DATABASE_ROUTES=[TestRouter]) # added router
def natural_key_test_multidb(self, format):
book1 = {
"data": "978-1590597255",
"title": "The Definitive Guide to Django: Web Development Done
Right",
}
book2 = {"data": "978-1590599969", "title": "Practical Django
Projects"}
# Create the books.
adrian = NaturalKeyAnchor.objects.create(**book1)
james = NaturalKeyAnchor.objects.create(**book2)
# Serialize the books.
string_data = serializers.serialize(
format,
NaturalKeyAnchor.objects.all(),
indent=2,
use_natural_foreign_keys=True,
use_natural_primary_keys=True,
)
# Delete one book (to prove that the natural key generation will only
# restore the primary keys of books found in the database via the
# get_natural_key manager method).
james.delete()
# Deserialize and test.
books = list(serializers.deserialize(format, string_data,
using='other')) # added using
self.assertCountEqual(
[(book.object.title, book.object.pk) for book in books],
[
(book1["title"], None), # None i.s.o. adrian.pk
(book2["title"], None),
],
)
}}}
It feels natural to me that `None` is returned when using a different
database in which the object doesn't exist. This test succeeds already
without the patch. I guess what's missing is that the router is not used
when setting up the database, but I'm currently somewhat lost in getting
it to do that. Any hints?
I'm going to look into it some more tomorrow.
--
Ticket URL: <https://code.djangoproject.com/ticket/34907#comment:3>
Django <https://code.djangoproject.com/>
The Web framework for perfectionists with deadlines.
--
You received this message because you are subscribed to the Google Groups
"Django updates" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/django-updates/0107018b6215076f-977cd934-ef7f-4be7-adc3-e2f5e90e4cff-000000%40eu-central-1.amazonses.com.
