When you create database unit test project, you need to setup connections to your database. I always use only one connection to run the code and test conditions. But I was curious what is the second connection for? Where does that validation occur?
Now I’ve spent some time to check when this validation connection is really used. I’ve created a simple unit test and placed a simple command to check user name in every part of unit test:
|
1 |
SELECT SUSER_NAME(); |
Here are the results of my checks.
| Part of the unit test | Connection used |
| Test initialize | Connection 2 |
| Pre-test | Connection 2 |
| Test execution | Connection 1 |
| Post-test | Connection 2 |
| Test cleanup | Connection 2 |
As you see, only test execution phase uses connection 1 (“execution”), whereas all the other phases use connection 2 (“validation”) if you use 2 connections. If you use only one connection, so all the steps will be run under this unique username.
It’s a bit strange that SSDT uses a term “validation” for a connection for pre- and post-test steps. But that’s out of our will 🙂