A scheam-first postgres cli migration tool.

go install github.com/aecsar/drift@latest

drift --help

The way I like to work when it comes to database migrations is to have a single schema.sql file that represents the current state of my database. I would then update that file and run a tool to diff it against the result execution of all pre-existing migrations. This approach presents some advantages, mainly enabling you to see the databse state without needing to open a connection or trying to parse existing migrations. The issue with ariga/atlas is it don't generate some schema objects (or require pro account?!) like triggers and functions and is simply an overkill sometimes.

drift -name "add_users_table"

# -migrations : migration directory, default to "./migrations/"
# -schema : schema file, default to "./schema.sql"

Drift uses stripe/pg-schema-diff under the hood. It applies all existing migrations to a temporary database using testcontainers. This mean you need to have docker to use drift. pg-schema-diff is then used to generate the diff migration by taking the schema content as target and temporary database as source.

Of course your schema file cannot live in the migrations directory simply because it will be treated as a migration.

• add tests