About the next spring a new auto incrementing primary keys features is expected to be released in Spanner. It should make autofield patch unnecessary at all.

+1. what is current status?
I've found the auto incrementing primary keys feature was introduced.
https://cloud.google.com/blog/products/databases/announcing-support-for-auto-generated-keys-in-spanner?hl=en

But still I'm facing error when migration info is inserted to django_migrations.

"id must not be NULL in table django_migrations."

On Cloud Spanner, following table is created automatically.
The table definition of django_migrations does not include configs for auto-generate-keys.

CREATE TABLE django_migrations (
  id INT64 NOT NULL,
  app STRING(255) NOT NULL,
  name STRING(255) NOT NULL,
  applied TIMESTAMP NOT NULL,
) PRIMARY KEY(id);;

I manually configured django_migrations with sequence like below and migration succeeded.

CREATE SEQUENCE serial OPTIONS(sequence_kind='bit_reversed_positive');

CREATE TABLE django_migrations (
  id INT64 DEFAULT (GET_NEXT_SEQUENCE_VALUE(SEQUENCE serial)),
  app STRING(255) NOT NULL,
  name STRING(255) NOT NULL,
  applied TIMESTAMP NOT NULL,
) PRIMARY KEY(id);;

but I'm concerning this is right way for work around. It would be helpful if someone give advice to me and for this workaround.

Facing this same issue - using spanner as a non-default DB. Whenever I am inserting records to another db - bulk_create is failing with thw following strace

File "/Users/saraf/Library/Caches/pypoetry/virtualenvs/workflow-platform-HlqtOrhH-py3.12/lib/python3.12/site-packages/django/contrib/contenttypes/management/__init__.py", line 152, in create_contenttypes ContentType.objects.using(using).bulk_create(cts) File "/Users/saraf/Library/Caches/pypoetry/virtualenvs/workflow-platform-HlqtOrhH-py3.12/lib/python3.12/site-packages/django/db/models/query.py", line 816, in bulk_create assert len(returned_columns) == len(objs_without_pk)

If I remove django_spanner from installed_apps before running this command - it works