fail
- Michael Bruton (Deactivated)
Draft in Progress
This document is a draft and is under development.
Description
Force a failure in an ETLUnit test. Useful for debugging.
Attributes
List of fail() Attributes
message: - >> TYPE: string - - REQUIRED
failure-id: - >> TYPE: string
Individual fail() Attributes
message:
- The message to send with the failure.
- This message will appear in the test log.
failure-id:
- The
failure-idstring is arbitrary. It may show up in the CLI (Command Line Interface) when you run the test to point to where the failure occurred. It may also show up on the summary page report next to the test method. - The convention is to make the failure-id value all capitals, with words separated by underscores. For instance,
FORCED_FAILUREorWORKFLOW_SHOULD_HAVE_FAILED.
Examples
fail() Example
When the etlunit test reaches the fail() operation, the test is marked for failure. However, the test method may continue processing code that appears after the fail() operation, until it reaches the end of the test method's code. It will then fail according to the attributes in the fail() operation.
@Test()
failOnPurpose()
{
log(
message: 'failing by operation fail()',
log-file-name: 'fail_me.log',
log-classifier: 'anyClassifier'
);
fail(
message:'Failure message 123-765',
failure-id: 'DELIBERATE_FAILURE'
);
}
The failure-id may show up in the CLI when you run the test.
userme > te #failOnPurpose
Processing [1] tests
class experiment.fail_operation ------------------------------------------------
1/1 .failOnPurpose
DELIBERATE_FAILURE
Failed F[1]
Tests run: 1, Failures: 1, Time elapsed: 00.269 sec
Note that the failure ID specified in the failure() operation appeared in the CLI under the name of the test method.
Run the report...
userme > r
. . . and you may see the value of error-id next to the method name.
Catch the failure generated by the fail() operation
The failure generated by the fail() operation may be caught by specifying the value of the failure-id attribute in the @Test annotation's expected-failure-id attribute.
@Test(expected-failure-id: 'DELIBERATE_FAILURE')
succeedWhilefailing()
{
log(
message: 'failing by operation fail()',
log-file-name: 'log_my_fail.log',
log-classifier: 'anyClassifier'
);
fail(
message:'Failure message 123-765',
failure-id: 'DELIBERATE_FAILURE'
);
}
The content of the test method hasn't changed, but because we specified an expected-failure-id, this time the test passed when run from the Command Line Interface:
userme > te #succeedWhile Processing [1] tests class experiment.fail_operation ------------------------------------------------ 1/1 .succeedWhilefailing Passed P[1] Tests run: 1, Successes: 1, Time elapsed: 00.179 sec
Using multiple fail() operations
@Test
failTwice()
{
fail(
message:'Failure message 123-765',
failure-id: 'DELIBERATE_FAILURE'
);
log(
message: 'failing by operation fail()',
log-file-name: 'fail_me.log',
log-classifier: 'anyClassifier'
);
fail(
message:'Failure message Whadduhyahknow',
failure-id: 'SECONDARY_FAILURE'
);
log(
message: 'I log again.',
log-file-name: 'fail_me_also.log',
log-classifier: 'anyClassifier'
);
}
Both failure IDs may show up on the CLI:
userme > te #failTwice
Processing [1] tests
class experiment.fail_operation ------------------------------------------------
1/1 .failTwice
DELIBERATE_FAILURE
SECONDARY_FAILURE
Failed F[2]
Tests run: 1, Failures: 2, Time elapsed: 00.181 sec
... and note that despite the fail() operation, the other code in the method went ahead and executed anyway. Check out the two logging messages in the test's log.
As further proof, the two log files were created when running this test method:
