SQLAlchemy 2.0 迁移 - 连接到数据库

Question

I am just testing out the SQLAlchemy 1.4 library out and I noticed in the foreseeable future they changing certain core functionalities. I generally connect to a database shown below, moving forward they mentioned they are going to remove this bind method on the MetaBase class, what is going to be the set standard of writing this?

def output_log(msg):
    connector = create_engine("mylogin")
    metadata = MetaData(bind=connector)
    log = Table("mytable", metadata, autoload=True)
    sess = connector.connect()
    i = insert(log).values({"Message": msg})
    sess.execute(i)
    sess.close()

Answer 1

The changes and background for this are described in “Implicit” and “Connectionless” execution, “bound metadata” removed .

You can rewrite your code as follows:

def output_log(msg):
    engine = create_engine("mylogin")
    metadata = MetaData()
    log = Table("mytable", metadata, autoload_with=engine)
    sess = engine.connect()
    i = log.insert().values({"Message": msg})
    sess.execute(i)
    sess.commit()
    sess.close()

Or for completeness using with :

    with engine.begin() as sess:
        i = log.insert().values({"Message": msg})
        sess.execute(i)

And the other half option with the future commit :

    with engine.connect() as sess:
        i = log.insert().values({"Message": msg})
        sess.execute(i)
        sess.commit()