Tips and tricks

Here you will find help on performing some common tasks related to SQLAlchemy with Asphalt.

How to automatically print emitted SQL

This can be done two ways:

  1. Add the echo=True option to the engine configuration options
  2. In your application’s configuration, add a logger for sqlalchemy.engine.base:
      handlers: [console]
      level: WARNING
      level: INFO
      level: INFO

Handling schema migrations

For schema migrations, it is best to use the Alembic tool, which is made by SQLAlchemy’s author. An ideal place to put your Alembic migration code in the application is the start() method of your application component, after calling await super().start(ctx), but before starting any services that might actually use the database. If this is not feasible, consider fishing the connection URL(s) out of self.component_configs dictionary and running the migration before calling the superclass start() method.

Assuming that you have an alembic directory in the same directory as the module containing the application component class, here’s how you might do it:

import os

from alembic import command, config

class ApplicationComponent(ContainerComponent):
    async def start(self, ctx):
        await super().start(ctx)

        cfg = config.Config(os.path.dirname(__file__), 'alembic', 'alembic.ini')
        with ctx.sql.bind.begin() as connection:
            cfg.attributes['connection'] = connection
            command.upgrade(cfg, "head")

Notice the direct use of the engine here – it’s okay as long as the connection created is short lived, as is guaranteed by doing with engine.begin():.