= Using Alembic for schema migrations =

To use Alembic for your next schema change:

Enable your virtualenv.

Create an initial database using setupdb_app.py. See the Makefile for examples of usage.

Make a change to:

    socorro/external/postgresql/models.py

And then:

    PYTHONPATH=. alembic revision --autogenerate -m "Making an important change to the schema"

A migration script will be put into:

    alembic/versions

Now you can modify it to meet your needs.

Then, to apply the migration:

    PYTHONPATH=. alembic upgrade heads

Downgrade to previous revision:

    PYTHONPATH=. alembic downgrade -1


== Dealing with unsupported or custom TYPEs ==

We use a few types not currently supported by SQLAlchemy, so you may
need to modify the migration slightly.


== Helper functions in socorro.lib.migrations

We have a couple helper functions for doing cleanup tasks, and adding custom
types. These functions and classes are imported by the default template for
a migration.

=== Fixing permissions after adding a table

This is hardcoded to change the owner of a table to `breakpad_rw`.

Example:

    fix_permissions(op, u'tablename')

=== Loading a new stored proc

Example:

    load_stored_proc(op, ['update_matviews.sql', 'backfill_matviews.sql'])

=== JSON datatype example

Modify to import JSON type from socorro's model:

    sa.Column(u'raw_crash', jsontype.JsonType(), nullable=False),

=== CITEXT datatype example

    sa.Column(u'new_stuff', citexttype.CitextType(), nullable=False),


== Adding an index

    sa.Column(u'uuid', postgresql.UUID(), nullable=False, index=True, unique=True),