kvesteri / sqlalchemy-continuum

Versioning extension for SQLAlchemy.
BSD 3-Clause "New" or "Revised" License
580 stars 126 forks source link

Using association-tables with multiple flush gives UNIQUE constraint error #296

Open AbdealiLoKo opened 2 years ago

AbdealiLoKo commented 2 years ago

When I have an association table, and I: Remove a value > flush > Add it back - sqla-continuum throws an error. But sqlalchemy handles it fine.

I have a simple example below. In my real application, I am calling various functions (Each doing their own flush() to ensure DB constraints are valid) And sometimes when they are trying to figure out valid/invalid values - they remove and re-add the associated relationship

Reproducible example:

from sqlalchemy import Column, ForeignKey, Integer, String, Table, create_engine
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy.orm import configure_mappers, relationship, sessionmaker
from sqlalchemy_continuum import make_versioned

make_versioned(user_cls=None)
Base = declarative_base()

class BaseModel(Base):
    __abstract__ = True

    def __repr__(self):
        return super().__repr__() + f" ({self.name})"

book_author = Table(
    "book_author",
    Base.metadata,
    Column("author_id", ForeignKey("author.id"), primary_key=True),
    Column("book_id", ForeignKey("book.id"), primary_key=True),
)

class Author(BaseModel):
    __tablename__ = "author"
    __versioned__ = {}
    id = Column(Integer, primary_key=True)
    name = Column(String(255))

    books = relationship("Book", secondary=book_author, back_populates="authors", lazy='selectin')

class Book(BaseModel):
    __tablename__ = "book"
    __versioned__ = {}
    id = Column(Integer, primary_key=True)
    name = Column(String(255))

    authors = relationship("Author", secondary=book_author, back_populates="books", lazy='selectin')

configure_mappers()
engine = create_engine('sqlite:////tmp/db.sqlite3', echo=True)

# Create session
session = sessionmaker()
session.configure(bind=engine)

# Create all models we need
Base.metadata.create_all(engine)

db = session()

adams = Author(name='Douglas Adams')
tolkien = Author(name='Tolkien')
db.add(adams)
db.add(tolkien)
db.flush()
print("FLUSH - 1")

h2g2 = Book(name='h2g2', authors=[tolkien])
db.add(h2g2)
db.flush()

h2g2.authors = [adams]  # Remove tolkien and add it back in another flush
db.flush()

h2g2.authors = [tolkien]
db.flush()

db.commit()