Main Content


Class: matlab.unittest.qualifications.Verifiable
Package: matlab.unittest.qualifications

Verify function throws specified exception


[output1,...,outputN] = verifyError(___)


verifyError(verifiable,actual,identifier) verifies that actual is a function handle that throws an exception with an error identifier that is equal to identifier.

verifyError(verifiable,actual,metaClass) verifies that actual is a function handle that throws an exception whose type is defined by the meta.class instance specified in metaClass. This method does not require the instance to be an exact class match, but rather it must be in the specified class hierarchy, and that hierarchy must include the MException class.

verifyError(___,diagnostic) also associates the diagnostic information in diagnostic with the qualification. Depending on the test runner configuration, the testing framework might display diagnostics when the qualification passes or fails. By default, the framework displays diagnostics only when the qualification fails. You can override the default behavior by customizing the test runner. For example, use a DiagnosticsOutputPlugin instance to display both failing and passing event diagnostics.

[output1,...,outputN] = verifyError(___) returns multiple output arguments from the invocation of the function handle actual. Use this syntax to control the number of output arguments requested from the function handle. If the function handle errors, all outputs are of type missing. Otherwise, output1,...,outputN are output values from actual. You can use this syntax with any of the input arguments of the previous syntaxes.

Input Arguments


The matlab.unittest.TestCase instance which is used to pass or fail the verification in conjunction with the test running framework.


The value to test.


Error identifier, specified as a character vector.


An instance of meta.class.


Diagnostic information related to the qualification, specified as one of the following:

  • string array

  • character array

  • function handle

  • matlab.unittest.diagnostics.Diagnostic object

Diagnostic values can be nonscalar. For more information, see matlab.unittest.diagnostics.Diagnostic.


expand all

testCase = matlab.unittest.TestCase.forInteractiveUse;

% Passing scenarios
verifyError(testCase, @() error('SOME:error:id','Error!'), 'SOME:error:id');
verifyError(testCase, @testCase.assertFail, ...
% Failing scenarios
verifyError(testCase, 5, 'some:id', '5 is not a function handle');
verifyError(testCase, @testCase.verifyFail, ...
    ?matlab.unittest.qualifications.AssertionFailedException, ...
    'Verifications dont throw exceptions.');
verifyError(testCase, @() error('SOME:id'), 'OTHER:id', 'Wrong id');
verifyError(testCase, @() error('whoops'), ...
    ?matlab.unittest.qualifications.AssertionFailedException, ...
    'Wrong type of exception thrown');

Create testNonNumericInput to test if function throws expected error message, add5:InputMustBeNumeric, for unexpected condition, input is char.

Function for unit testing:

function res = add5(x)
% ADD5 Increment input by 5.
if ~isa(x,'numeric')
    error('add5:InputMustBeNumeric','Input must be numeric.')
res = x + 5;

TestCase class containing test methods:

classdef Add5Test < matlab.unittest.TestCase
    methods (Test)
        function testDoubleOut(testCase)
            actOutput = add5(1);
        function testNonNumericInput(testCase)

Create a test suite from the Add5Test class file.

suite = matlab.unittest.TestSuite.fromFile('Add5Test.m')
result = run(suite);
Running Add5Test
Done Add5Test

In your current working folder, create the following function to test. The function returns a single output.

function y = plusOne(x)
    y = x+1;

At the command prompt, call plusOne with too many output arguments.

[~,~] = plusOne(41);
Error using plusOne
Too many output arguments.

View the error message and identifier.

ans = 

  MException with properties:

    identifier: 'MATLAB:TooManyOutputs'
       message: 'Too many output arguments.'
         cause: {0×1 cell}
         stack: [0×1 struct]

Verify that if the plusOne function is called with too many requested output arguments, it errors and the error has the identifier MATLAB:TooManyOutputs.

testCase = matlab.unittest.TestCase.forInteractiveUse;
[~,~] = verifyError(testCase,@() plusOne(41),'MATLAB:TooManyOutputs');
Interactive verification passed.


  • This method is functionally equivalent to:

    import matlab.unittest.constraints.Throws;
    verifiable.verifyThat(actual, Throws(identifier));
    verifiable.verifyThat(actual, Throws(metaClass));

    There exists more functionality when using the Throws constraint directly via verifyThat.

  • Use verification qualifications to produce and record failures without throwing an exception. Since verifications do not throw exceptions, all test content runs to completion even when verification failures occur. Typically verifications are the primary qualification for a unit test since they typically do not require an early exit from the test. Use other qualification types to test for violation of preconditions or incorrect test setup. Alternatively,

    • Use assumption qualifications to ensure that the test environment meets preconditions that otherwise do not result in a test failure. Assumption failures result in filtered tests, and the testing framework marks the tests as Incomplete. For more information, see matlab.unittest.qualifications.Assumable.

    • Use assertion qualifications when the failure condition invalidates the remainder of the current test content, but does not prevent proper execution of subsequent test methods. A failure at the assertion point renders the current test method as failed and incomplete. For more information, see matlab.unittest.qualifications.Assertable.

    • Use fatal assertion qualifications to abort the test session upon failure. These qualifications are useful when the failure mode is so fundamental that there is no point in continuing testing. These qualifications are also useful when fixture teardown does not restore the MATLAB® state correctly and it is preferable to abort testing and start a fresh session. For more information, see matlab.unittest.qualifications.FatalAssertable.

Introduced in R2013a