Want to use custom schema name for created database tables

[kaysweebe]

TLDR; Using custom schema name does not allow me to insert without getting operational errors because the insert tries to use the default schema name given on the model object.

Example: CrontabSchedule.table_args.schema = 'celery_schema'

Currently, the only schema name supported is celery_schema. I can add these tables super efficiently into my database with schema my_example_schema so that in my database it shows my_example_schema.celery_crontabschedule using the normal start up process; however, the celery_schema name is hard-coded on each of the models making it really difficult to use a custom schema name.

Using MySQL by the way

[kaysweebe]

I actually have a fix for this now. Not too sure if you allow others to open PR's on this repo?

[farahats9]

Any PR is welcome for sure.

But did you try to use the beat_schema config variable in the celery beat config dict?
it would looks something like this:

config = {
    'beat_schedule': {
        "run_every_hour": {
            "task": "tasks.run_every_hour",
            "schedule": schedules.crontab(hour="*", minute=0),
        },
    },
    'beat_dburi': 'mysql+mysqldb://user:password@127.0.0.1:3306/test',
    'beat_schema': 'my_example_schema',
    ...
}

celery.conf.update(config)

I don't remember if it works well on MySQL or if the database ignores it, but the first run of celery beat it should create the tables in the schema you described.
If you have a better solution please go ahead and open a PR.