ODBC 作为一项技术存在什么问题？
最近 Zed Shaw（一位博客程序员）提到应该从流行的 Python 书籍 Dive into Python 中删除 ODBC 引用。我从未使用过 ODBC，我只是想了解为什么 ODBC 如此“糟糕”。该技术的优点和缺点是什么？还有哪些替代方案？
Recently Zed Shaw (a programmer who blogs) mentioned that ODBC references should be removed from the popular python book Dive into Python. I have never worked with ODBC and I just wanted to understand why ODBC is so "bad". What are the pros and cons of the technology? What alternatives are there?
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（3）
最大的问题是 64 位支持并不完全适用于不同的数据源。另一个问题是，人们经常将 ODBC 与另一种技术桥接起来，这样他们就可以简单地消除不必要的复杂层。例如，当他们可以直接直接使用 JDBC 时，JDBC-> ODBC。
The biggest issue is that 64 Bit support is not entirely available across different data sources. Another problem is that people often bridge ODBC from another technology where they could simply eliminate the unnecessary layers of complexity. For example, JDBC->ODBC when they could simply go JDBC direct.
ODBC 是我发现的从 Python 访问 Solid DB 服务器的唯一方法。
如果您别无选择，它可能会成为您的救星。
ODBC is the only way I've found to access a Solid DB Server from Python.
If you have no other choice, it could be a life saver.
只是补充一下 Cat 的评论：似乎许多 JDBC 驱动程序都是 ODBC 代码的非常薄的包装器，或者，特定的 ODBC 和 JDBC 驱动程序可能是其“下面”的公共代码的薄包装器。
曾经有一段时间 ODBC 也很慢/有缺陷。它名声不佳，MS 坚持每 n 年引入一种新的数据访问技术，这意味着他们的 ODBC 驱动程序没有改进，而他们的新系统复制粘贴了许多错误（就像在糟糕的文本驱动程序）。然而，那些没有其他系统的人，比如使用 iODBC 的人，插入并最终获得了一套强大、可靠且快速的驱动程序。任何早期对性能的抱怨早已得到解决。
今天我对 ODBC 的主要抱怨是驱动程序安装开销。 JDBC 可以即时完成此操作，这对于从用户到 IT 部门的每个人来说都更容易。
Just to add to Cat's comment: it seems that many JDBC drivers are very thin wrappers over ODBC code, or alternately, that a particular ODBC and JDBC driver may be thin wrappers over common code "below" it.
There was a time when ODBC was also slow/buggy. It gained a bad reputation, and MS's insistence on introducing a new data access technology every n years meant that their ODBC drivers didn't improve, while their new system copy-n-pasted many of the bugs over (as is the case in the gawd-awful text driver). However, those who had no other system, like the iODBC guys, plugged on and ended up with a suite of powerful, solid and fast drivers. Any early complaints about performance were addressed long-ago.
My main complaint with ODBC today is the driver installation overhead. JDBC can do that on-the-fly, which is much easier on everyone from the user to the IT department.