哦了测试存储过程和SQL查询你需要：

• 搭建 postgres 环境.
• 创建一个测试服务器.
• 测试环境需要设置
• 所有的测试套件需要测试和Rollbacked（关键）。
• 测试数据需要添加. YAAAAAWNNN!
• CI 集成, Docker. YAAAAAAAAWNNNN!

这就是所需要的大量工作！幸运的是pypgTAP能够解决了很多这些问题。

• Oh And did we mention its in python ?

它是如何做到的呢？

• 它启动一个使用后可抛弃的postgres服务器，来执行你的测试。然后并丢弃服务器和它的文件。
• pypgTAP可以在你的Postgres的代码库中，以一种相互隔离的方式来运行多个测试套件。
• Virtualenv python support: The industry standard for using python is virtualenv and tox and its a pain to use pgTAP with python. Remember plpython procedures execute on the postgres server and it cannot find python procedures executed by a client like psql Example:

CREATE OR REPLACE FUNCTION somefunction(arg1 varchar, arg2 varcha...)  RETURNS varchar  AS $$     print 'Hello World'     import myutility -- import error if myutility is not found in global site-packages    $$ LANGUAGE plpythonu strict;

• The virtualenv sharing also facilitates reuse of existing python code base, And hey we all just want an excuse to write in pure python :)
• It also makes it easy for people to do SDLC in pure python when its necessary and just use it seamlessly from within postgres's plpythonu.
• You can refer to resources within your project root in your unit tests and pypgtap will find them for you.
• You can test standalone SQL code (Like for Postgres Derivatives like redshift that don't support stored procedures) or test stored procedures.

项目主页：http://www.open-open.com/lib/view/home/1418787193058