我一直在尝试使用夹具将测试数据集加载到我的 Pylons / PostgreSQL 应用程序中。这很好用,但如果它们引用自动增量 id 字段,则无法正确创建外键。
我的夹具看起来像这样:
class AuthorsData(DataSet):
class frank_herbert:
first_name = "Frank"
last_name = "Herbert"
class BooksData(DataSet):
class dune:
title = "Dune"
author_id = AuthorsData.frank_herbert.ref('id')
和模型:
t_authors = sa.Table("authors", meta.metadata,
sa.Column("id", sa.types.Integer, primary_key=True),
sa.Column("first_name", sa.types.String(100)),
sa.Column("last_name", sa.types.String(100)),
)
t_books = sa.Table("books", meta.metadata,
sa.Column("id", sa.types.Integer, primary_key=True),
sa.Column("title", sa.types.String(100)),
sa.Column("author_id", sa.types.Integer, sa.ForeignKey('authors.id'))
)
运行“paster setup-app development.ini”时,SQLAlchemey 将 FK 值报告为“None”,因此显然没有找到它:
15:24:27,284 INFO [sqlalchemy.engine.base.Engine.0x...9f50]
CREATE TABLE authors (
id SERIAL NOT NULL,
first_name VARCHAR(100),
last_name VARCHAR(100),
PRIMARY KEY (id)
)
15:24:27,284 INFO [sqlalchemy.engine.base.Engine.0x...9f50] {}
15:24:27,291 INFO [sqlalchemy.engine.base.Engine.0x...9f50] COMMIT
15:24:27,292 INFO [sqlalchemy.engine.base.Engine.0x...9f50]
CREATE TABLE books (
id SERIAL NOT NULL,
title VARCHAR(100),
author_id INTEGER,
PRIMARY KEY (id),
FOREIGN KEY(author_id) REFERENCES authors (id)
)
15:24:27,293 INFO [sqlalchemy.engine.base.Engine.0x...9f50] {}
15:24:27,300 INFO [sqlalchemy.engine.base.Engine.0x...9f50] COMMIT
15:24:27,301 INFO [blog.websetup] Inserting initial data
15:24:27,302 INFO [sqlalchemy.engine.base.Engine.0x...9f50] BEGIN
15:24:27,309 INFO [sqlalchemy.engine.base.Engine.0x...9f50] INSERT INTO authors (first_name, last_name) VALUES (%(first_name)s, %(last_name)s) RETURNING authors.id
15:24:27,309 INFO [sqlalchemy.engine.base.Engine.0x...9f50] {'first_name': 'Frank', 'last_name': 'Herbert'}
15:24:27,320 INFO [sqlalchemy.engine.base.Engine.0x...9f50] INSERT INTO books (title, author_id) VALUES (%(title)s, %(author_id)s) RETURNING books.id
15:24:27,320 INFO [sqlalchemy.engine.base.Engine.0x...9f50] {'author_id': None, 'title': 'Dune'}
15:24:27,322 INFO [sqlalchemy.engine.base.Engine.0x...9f50] COMMIT
15:24:27,323 INFO [blog.websetup] Done
夹具文档实际上警告这可能是一个问题:
“但是,在某些情况下,您可能需要引用一个在加载之前没有值的属性,例如序列 ID 列。(请注意,使用会话时 SQLAlchemy 数据层不支持这一点。)”
http://farmdev.com/projects/fixture/using-dataset.html#referencing-foreign-dataset-classes
这是否意味着 SQLAlchemy 不支持这一点?或者是否可以在不使用 SA“会话”的情况下加载数据?其他人如何处理这个问题?