Using Transact-SQL Assertions in SQL Server Unit Tests


In a SQL Server unit test, a Transact-SQL test script runs and returns a result. Sometimes, the results are returned as a results set. You can validate results by using test conditions. For example, you can use a test condition to check how many rows were returned in a specific result set or to verify how long a particular test took to run. For more information about test conditions, see Using Test Conditions in SQL Server Unit Tests.

Instead of using test conditions, you can also use Transact-SQL assertions, which are THROW or RAISERROR statements in a Transact-SQL script. In certain circumstances, you might prefer to use a Transact-SQL assertion instead of a test condition.

You should consider the following points before you decide to validate data either by using Transact-SQL assertions or by using test conditions.

  • Performance. It is faster to run a Transact-SQL assertion on the server than to first move data to a client computer and manipulate it locally.

  • Familiarity with language. You might prefer a particular language based on your current expertise and therefore choose Transact-SQL assertions or Visual C# or Visual Basic test conditions.

  • Complicated validation. In some instances, you can build more complex tests in Visual C# or Visual Basic and validate your tests on the client.

  • Simplicity. It is often simpler to use a pre-defined test condition than to write the equivalent script in Transact-SQL.

  • Legacy validation libraries. If you already have code that performs validation, you can use it in a SQL Server unit test instead of using test conditions.

To mark a SQL Server unit test method with expected exceptions, add the following attribute:

[ExpectedSqlException(MessageNumber=nnnnn, Severity=x, MatchFirstError=false, State=y)]


  • nnnnn is the number of the expected message, for example 14025

  • x is the severity of the expected exception

  • y is the state of the expected exception

Any unspecified parameters are ignored. You pass these parameters to the RAISERROR statement in your database code. If you specify MatchFirstError = true, the attribute will match any of the SqlErrors in the exception. The default behavior (MatchFirstError = true) is to only match the first error that occurs.

For an example of how to use expected exceptions and a negative SQL Server unit test, see Walkthrough: Creating and Running a SQL Server Unit Test.


Use THROW instead of RAISERROR. RAISERROR is now deprecated.

You can directly use Transact-SQL assertions on the server by using the RAISERROR statement in your Transact-SQL script. Its syntax is:

RAISERROR (@ErrorMessage, @ErrorSeverity, @ErrorState)


@ErrorMessage is any user-defined error message. You can format this message string similar to the printf_s function.

@ErrorSeverity is a user-defined severity level from 0 – 18.


The values '0' and '10' for the severity level do not cause the SQL Server unit test to fail. You can use any other value in the range 0 - 18 to cause the test to fail.

@ErrorState is an arbitrary integer from 1 – 127. You can use this integer to differentiate between occurrences of a single error that is raised at different locations in the code.

For more information, see RAISERROR (Transact-SQL). An example of using RAISERROR in a SQL Server unit test is provided in the topic, How to: Write a SQL Server Unit Test that Runs within the Scope of a Single Transaction.