Alembic tries to drop index when using TimescaleDB hypertable

[kimdre]

When I use TimescaleDB and create a hypertable on a normal PostgreSQL table, a new index on the given time column gets created automatically.
Since this index is not created from any migrations, Alembic tries to delete the index whenever I autogenerate a new migration.

Is there a way to tell Alembic to ignore this "external" index, e.g. by its name?

This is the migration content that always gets generated.

def upgrade() -> None:
    # ### commands auto generated by Alembic - please adjust! ###
    op.drop_index('events_timestamp_idx', table_name='events')
    # ### end Alembic commands ###


def downgrade() -> None:
    # ### commands auto generated by Alembic - please adjust! ###
    op.create_index('events_timestamp_idx', 'events', [sa.text('timestamp DESC')], unique=False)
    # ### end Alembic commands ###

[CaselIT]

Hi,

Sure, take a look at https://alembic.sqlalchemy.org/en/latest/api/runtime.html#alembic.runtime.environment.EnvironmentContext.configure.params.include_name

[kimdre]

Would this be the correct approach?

def exclude_by_name(name, type_, parent_names):
    exclude_indexes = ["events_timestamp_idx"]
    if type_ == "index":
        return name not in exclude_indexes
    else:
        return True

context.configure(
    # ...
    include_name=exclude_by_name,

[CaselIT]

It should be yes

[MRigal]

By using the TimescaleDB dialect from this library, you may not have to ignore it anymore... https://github.com/dorosch/sqlalchemy-timescaledb