Running tests

The utPLSQL framework provides two main entry points to run unit tests from within the database:

  • procedures and functions
  • procedures

These two entry points differ in purpose and behavior. Most of the time you will want to use as ut_runner is designed for API integration and does not output the results to the screen directly.

Running from CI servers and command line

The best way to run your tests from CI server or command line is to use the utPLSQL-cli command line client.

Amongst many benefits it provides ability to:

  • see the progress of test execution for long-running tests - real-time reporting
  • use many reporting formats simultaneously and save reports to files (publish)
  • map your project source files and test files into database objects

You may download the latest release of the command line client from here or do it automatically using the command below (Unix).

# Get the url to latest release "zip" file
DOWNLOAD_URL=$(curl --silent | awk '/zipball_url/ { print $2 }' | sed -r 's/"|,//g')
# Download the latest release "zip" file
curl -Lk "${DOWNLOAD_URL}" -o
# Extract downloaded "zip" file
unzip -q

The ut package contains overloaded run procedures and functions. The run API is designed to be called directly by a developer when using an IDE/SQL console to execute unit tests. The main benefit of using this API is it’s simplicity. A single line call is enough to execute a set of tests from one or more schemes.

The procedures execute the specified tests and produce output to DBMS_OUTPUT using the specified reporter. The functions can only be used in SELECT statements. They execute the specified tests and produce outputs as a pipelined data stream to be consumed by a select statement. procedures

The examples below illustrate different ways and options to invoke procedures.

alter session set current_schema=hr;

Executes all tests in current schema (HR).


Executes all tests in specified schema (HR).


Executes all tests from all packages that are on the com.my_org.my_project suitepath. Check the annotations documentation to find out about suitepaths and how they can be used to organize test packages for your project.


Executes all tests from package hr.test_apply_bonus.


Executes single test procedure hr.test_apply_bonus.bonus_cannot_be_negative.


Executes all tests from package hr.test_apply_bonus and all tests from schema cust.


Executes all tests from package hr.test_apply_bonus and all tests from schema cust.


Executes all tests from package hr.test_apply_bonus and all tests from schema cust.

Using a list of items to execute allows you to execute a fine-grained set of tests.

List can be passed as a comma separated list or a list of ut_varchar2_list objects or as a list within ut_varchar2_list.


ut_documentation_reporter is the default reporter for all APIs defined for running unit tests.

The procedures and functions accept a_reporter attribute that defines the reporter to be used in the run. You can execute any set of tests with any of the predefined reporters.

begin'hr.test_apply_bonus', ut_xunit_reporter());

Executes all tests from package HR.TEST_APPLY_BONUS and provide outputs to DBMS_OUTPUT using the XUnit reporter.

For details on build-in reporters look at reporters documentation. functions

The functions provide exactly the same functionality as the procedures. You may use the same sets of parameters with both functions and procedures. The only difference is the output of the results. Functions provide output as a pipelined stream and therefore need to be executed as select statements.


select * from table('hr.test_apply_bonus', ut_xunit_reporter())); procedures

The ut_runner package provides an API for integrating utPLSQL with other products. Maven, Jenkins, SQL Develper, PL/SQL Developer, TOAD and others can leverage this API to call utPLSQL.

The main difference compared to the API is that does not print output to the screen. accepts multiple reporters. Each reporter pipes to a separate output (uniquely identified by output_id). Outputs of multiple reporters can be consumed in parallel. This allows for live reporting of test execution progress with threads and several database sessions.

The concept is pretty simple.

  • in the main thread (session), define the reporters to be used. Each reporter has it’s output_id and so you need to extract and store those output_ids.
  • as a separate thread, start and pass reporters with previously defined output_ids.
  • for each reporter start a separate thread and read outputs from the ut_output_buffer.get_lines table function by providing the output_id defined in the main thread.