Locking in sqlalchemy

Question:

I’m confused about how to concurrently modify a table from several different processes. I’ve tried using Query.with_lockmode(), but it doesn’t seem to be doing what I expect it to do, which would be to prevent two processes from simultaneously querying the same rows. Here’s what I’ve tried:

import time
from sqlalchemy.orm import sessionmaker, scoped_session
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy import *

engine = create_engine('mysql://...?charset=utf8&use_unicode=0', pool_recycle=3600, echo=False)
Base = declarative_base(bind=engine)
session = scoped_session(sessionmaker(engine))

class Test(Base):
    __tablename__ = "TESTXYZ"
    id = Column(Integer, primary_key=True)
    x = Column(Integer)

def keepUpdating():
    test = session.query(Test).filter(Test.id==1).with_lockmode("update").one()

    for counter in range(5):
        test.x += 10
        print test.x
        time.sleep(2)

    session.commit()


keepUpdating()

If I run this script twice simultaneously, I get session.query(Test).filter(Test.id==1).one().x equal to 50, rather than 100 (assuming it was 0 to begin with), which was what I was hoping for. How do I get both processes to either simultaneously update the values or have the second one wait until the first is done?

Asked By: Noah

||

Answers:

Are you by accident using MyISAM tables? This works fine with InnoDB tables, but would have the described behavior (silent failure to respect isolation) with MyISAM.

Answered By: Ants Aasma
Answered By: Alex_Dom
Categories: questions Tags: , , ,
Answers are sorted by their score. The answer accepted by the question owner as the best is marked with
at the top-right corner.