Question

I am using sql alchemy in my project, I used db session,

engine = create_engine(configuration)
db_session = scoped_session(sessionmaker(autocommit=False,
                                     autoflush=False,
                                     bind=engine))

Base = declarative_base()
Base.query = db_session.query_property()

def init_db():    
    import models
    Base.metadata.create_all(bind=engine)

DB session used as:

db_session.merge(order)      #order(model) in object
db_session.commit()

Now I want to insert data in to two tables order and order line item, so I need transaction,
as:
1. In first insert I want inserted order’s id to use in second insert query
2. If second insert query failed then first query should be rollback

Try:
    #begin transaction/How to begin transaction?
    order=db_session.add(order)      #insert into order
    #is need to commit db_session here as I need inserted orders id
    #here actually db_session.commit() needed to get order's id(auto generated) 
    #if db_session committed here then sql alchemy starts new session       

    order_line_item.id = order.id
    db_session.add(order_line_item)    #insert into order line line item

    db_session.commit()
    #check transaction status if failed then rollback, How to check status?

except:
    db_session.rollback()

How to use trasaction?

Answer 1

Nested transaction, as suggested by Lafada, is not intended for such situation. A flush() will do just fine, e.g.

db_session.begin()
try:
    db_session.add(order)

    db_session.flush()

    order_line_item.id = order.id
    db_session.add(order_line_item)

    db_session.commit()
except:
    db_session.rollback()

Or better still, if you have relationship setup properly between order and line item, you shouldn’t even bother to assign id manually.