Sqlalchemy：基于两个键访问一个值

Question

I'm working on a program using sqlalchemy and want to have one table that stores all the data related to the program. Each user of the program has a unique user ID, and one or more items being tracked by the program. For each item, there is a single float. The database layer of the program should, given a user ID and an item name (both strings), provide access to the float or allow for it to be overwritten.

My understanding of the object-relational mapper is that it allows for rows in a table to behave like a storage location for objects on disk memory, but I may be incorrect. I have written the following code, but I'm at a loss for how to retrieve instances, modify them, or create new ones.

import datetime
import sqlalchemy as sql

engine = sql.create_engine('sqlite:///:memory:', echo=True)
Base = declarative_base()

class Item(Base):
    __tablename__ = 'items'
    account_id = sql.Column(sql.String, primary_key=True)
    name = sql.Column(sql.String, primary_key=True)
    timestamp = sql.Column(sql.Float)

    def set_now(self):
        """Sets self.timestamp to the Unix timestamp of the present time."""
        self.timestamp = datetime.datetime.now().timestamp()

    def is_today(self):
        """Returns True if self.timestamp is part of the current day"""
        timestamp_dt = datetime.datetime.fromtimestamp(self.timestamp)
        return timestamp_dt.date() == datetime.datetime.today().date()

Answer 1

You need to create a sessionmaker and a session to be able to save or retrieve the records. Of course you also need to create the table. For example:

Item.__table__.create(bind=engine)
Session = sessionmaker(bind=engine)
session = Session()

item = Item(account_id='a1', name='n1')
item.set_now()
session.add(item)
session.commit()