Question

我正在使用序列化程序转储和加载编写sqlalchemy导入/导出脚本。出口工作，但我有进口问题，主要是由于外键问题。我正在使用sorted_tables来获取基于依赖项排序的表的列表，这确保我不会有交叉表外键问题但是有类似处理内部外键（指向自身的表）吗？

我基本上在考虑两种可能的解决方案：

找到一种基于依赖关系对行进行排序的方法
禁用所有约束 - ＆gt;插入数据 - ＆gt;启用所有约束再次

但我不确定如何正确地做到这一点......

表格示例：

class Employee(Base):
    __tablename__ = "t_employee"

    id = sa.Column(Identifier, sa.Sequence('%s_id_seq' % __tablename__), primary_key=True, nullable=False)
    first_name = sa.Column(sa.String(30))
    last_name = sa.Column(sa.String(30))
    manager_id  = sa.Column(Identifier, sa.ForeignKey("t_employee.id", ondelete='SET NULL'))

这是我的剧本：

def export_db(tar_file):
    print "Exporting Database. This may take some time. Please wait ..."

    Base.metadata.create_all(engine)
    tables = Base.metadata.tables

    with tarfile.open(tar_file, "w:bz2") as tar:
        for tbl in tables:
            print "Exporting table %s ..." % tbl
            table_dump = dumps(engine.execute(tables[tbl].select()).fetchall())

            ti = tarfile.TarInfo(tbl)
            ti.size = len(table_dump)
            tar.addfile(ti, StringIO(table_dump))

    print "Database exported! Exiting!"

    exit(0)

def import_db(tar_file):
    print "Importing to Database. This may take some time. Please wait ..."

    print "Dropping all tables ..."
    Base.metadata.drop_all(engine)

    print "Creating all tables ..."
    Base.metadata.create_all(engine)
    tables = Base.metadata.sorted_tables

    with tarfile.open(tar_file, "r:bz2") as tar:
        for tbl in tables:
            try:
                entry = tar.getmember(tbl.name)
                print "Importing table %s ..." % entry.name
                fileobj = tar.extractfile(entry)
                table_dump = loads(fileobj.read(), Base.metadata, db)
                for data in table_dump:
                    db.execute(tbl.insert(), strip_unicode(dict(**data)))
            except:
                traceback.print_exc(file=sys.stdout)
                exit(0)
        db.commit()

    print "Database imported! Exiting!"

    exit(0)

Answer 1

对于批量转储，标准技术是禁用约束，执行导入，然后重新启用它们。您还可以在插入上获得更快的性能。

sqlalchemy导入导出IntegrityError失败

1 个答案: