evaulate support of schema_translate_map feature of SQLAlchemy with alembic migrations

Hello

I'm using the following version script for upgrade:

    op.create_table('feature_class',
        sa.Column('id', sa.Integer(), nullable=False),
        sa.Column('name', sa.String(length=1024), nullable=False),
        sa.PrimaryKeyConstraint('id'),
        sa.UniqueConstraint('name')
    )
    op.add_column(
        'features',
        sa.Column('feature_class_id', sa.Integer(), nullable=True),
    )

The table features still exists from a version before.
When running the upgrade it fails with the error:

sqlalchemy.exc.ProgrammingError: (psycopg2.ProgrammingError) relation "features" does not exist
 [SQL: 'ALTER TABLE features ADD COLUMN feature_class_id INTEGER'] (Background on this error at: http://sqlalche.me/e/f405)

I enabled SQLAlchemy SQL statement output and was able to see the following SQL statements for the operation:

CREATE TABLE develop_test.feature_class (
        id SERIAL NOT NULL,
        name VARCHAR(1024) NOT NULL,
        PRIMARY KEY (id),
        UNIQUE (name)
)

ALTER TABLE features ADD COLUMN feature_class_id INTEGER

As you can see the CREATE TABLE statement is using the schema name develop_test, but not the ALTER TABLE statement.

I apply the schema on this database by using the following SQLAlchemy command:

engine.update_execution_options(schema_translate_map={None: 'develop_test'})

Where engine is the engine object, which have been bound on the declarative class which I use for all table classes.

In my env.py script I use exactly this engine object to create the connection:

    # database is an "container"-like object, which collects several information about 
    # my database, like the engine, metadata and schema objects
    connectable = database.engine 

    with connectable.connect() as connection:
        context.configure(
            connection=connection,
            target_metadata=target_metadata,
            process_revision_directives=process_revision_directives,
            version_table_schema=database.schema 
        )

        with context.begin_transaction():
            context.run_migrations()

Maybe I'm using the schema in a wrong way?

Start in env.py at context.configure and the migration operations shown in the issue. Trace how the connection's schema_translate_map reaches run_migrations, comparing CREATE TABLE with op.add_column's generated ALTER TABLE. Reproduce the behavior against PostgreSQL and document or test the expected schema-qualified behavior.