To see this working, head to your live site.

Alembic migration: How to set server_onupdate in Alembic's alter_column function

I'm trying to change a table column in PostgreSQL using Alembic but I don't know how to perform the needed update to apply the SQLAlchemy's server_onupdate property.

The column is:

changed = Column(ArrowType(timezone=True), server_default=utcnow(), primary_key=True)

I'm using the Arrowtype column type from SQLAlchemy_utils package (this is not a problem).

My intention is to create something like this:

changed = Column(ArrowType(timezone=True), **server_onupdate=utcnow()**, primary_key=True)

But using the Alembic function: alter_column

In the documentation there are only references to the server_default property but nothing about server_onupdate

Is there a way to achieve this?

Thanks

1 comment

Answers (1)

AnswerDeveloper

Dec 28, 2022

To change the server_onupdate property of a column using Alembic, you can use the server_onupdate argument in the alter_column function, like this:


from alembic import op
from sqlalchemy import Column, ArrowType
from sqlalchemy_utils import utcnow

def upgrade():
    op.alter_column('table_name', 'changed',
                    server_onupdate=utcnow(),
                    existing_type=ArrowType(timezone=True),
                    existing_server_default=utcnow(),
                    existing_nullable=False)

This will change the server_onupdate property of the changed column to utcnow() in the table_name table.

Note that you will also need to provide the existing data type and server default value of the column using the existing_type and existing_server_default arguments, respectively. This is because Alembic needs to know the current state of the column in order to update it.