在 Flask/SQLAlchemy 中对 InstrumentedAttribute 选择性地执行数学计算

Question

I'm working with the Flask / SQLAlchemy ORM and what I'd like to do is to perform mathematical calculations selectively on instances of sqlalchemy.orm.attributes.InstrumentedAttribute . An equation like this would look something like User.query.update({User.score: User.points / User.time}) .

For example, let's say I have > 100,000 entries in my PostgreSQL database that looks something like this: User.points --> [12, 3, 0, ..., 45]; User.time --> [102, 23, 45, ..., 12] (please note the 0 in User.points ). In the example below, I would get an exception stating I cannot take the logarithm of 0.

from sqlalchemy import func
from project.models import User


User.query.update({User.score: func.log(User.points / User.time)})

I'd like to selectively choose which InstrumentAttribute to perform calculations on. I can use User.query.filter to filter a query, but how would I transform this into a filtered model without performing calculations iteratively, if you will?

Answer 1

db.session.query(User).where(User.points != 0).update({User.score: func.log(User.points / User.time)})
// or
User.query.filter(User.points != 0).update({User.score: func.log(User.points / User.time)})
db.session.commit()

logs:

2021-07-20 20:20:25,750 INFO sqlalchemy.engine.Engine BEGIN (implicit)
2021-07-20 20:20:25,754 INFO sqlalchemy.engine.Engine UPDATE "user" SET score=log("user".points / "user".time) WHERE "user".points != %(points_1)s
2021-07-20 20:20:25,754 INFO sqlalchemy.engine.Engine [no key 0.00046s] {'points_1': 0}
2021-07-20 20:20:25,764 INFO sqlalchemy.engine.Engine UPDATE "user" SET score=log("user".points / "user".time) WHERE "user".points != %(points_1)s
2021-07-20 20:20:25,764 INFO sqlalchemy.engine.Engine [no key 0.00055s] {'points_1': 0}
2021-07-20 20:20:25,766 INFO sqlalchemy.engine.Engine COMMIT