Boost C++ Libraries of the most highly regarded and expertly designed C++ library projects in the world. Herb Sutter and Andrei Alexandrescu, C++ Coding Standards

This is the documentation for an old version of Boost. Click here to view this page for the latest version.

To define properly datasets, the notion of sample should be introduced first. A sample is defined as polymorphic tuple. The size of the tuple will be by definition the arity of the sample itself.

A dataset is a collection of samples, that

Hence the dataset implements the notion of sequence.

The descriptive power of the datasets in Unit Test Framework comes from

[Tip] Tip

Only "monomorphic" datasets are supported, which means that all samples within a single dataset have the same type and same arity [2] . However, dataset of different sample types may be combined together with zip and cartesian product.

As we will see in the next sections, datasets representing collections of different types may be combined together (e.g.. zip or grid). These operations result in new datasets, in which the samples are of an augmented type.

The interface of the dataset should implement the two following functions/fields:

  • iterator begin() where iterator is a forward iterator,
  • boost::unit_test::data::size_t size() const indicates the size of the dataset. The returned type is a dedicated class size_t that can indicate an infinite dataset size.
  • an enum called arity indicating the arity of the samples returned by the dataset

Once a dataset class D is declared, it should be registered to the framework by specializing the template class


with the condition that


evaluates to true.

The following example implements a custom dataset generating a Fibonacci sequence.

Example: Example of custom dataset


#define BOOST_TEST_MODULE dataset_example68
#include <boost/test/included/unit_test.hpp>
#include <boost/test/data/test_case.hpp>
#include <boost/test/data/monomorphic.hpp>
#include <sstream>

namespace bdata = boost::unit_test::data;

// Dataset generating a Fibonacci sequence
class fibonacci_dataset {
    // the type of the samples is deduced
    enum { arity = 1 };

    struct iterator {

        iterator() : a(1), b(1) {}

        int operator*() const   { return b; }
        void operator++()
            a = a + b;
            std::swap(a, b);
        int a;
        int b; // b is the output

    fibonacci_dataset()             {}

    // size is infinite
    bdata::size_t   size() const    { return bdata::BOOST_TEST_DS_INFINITE_SIZE; }

    // iterator
    iterator        begin() const   { return iterator(); }

namespace boost { namespace unit_test { namespace data { namespace monomorphic {
  // registering fibonacci_dataset as a proper dataset
  template <>
  struct is_dataset<fibonacci_dataset> : boost::mpl::true_ {};

// Creating a test-driven dataset, the zip is for checking
    fibonacci_dataset() ^ bdata::make( { 1, 2, 3, 5, 8, 13, 21, 35, 56 } ),
    fib_sample, exp)
      BOOST_TEST(fib_sample == exp);


> example68
Running 9 test cases...
test.cpp(60): error: in "test1/_7": check fib_sample == exp has failed [34 != 35]
Failure occurred in a following context:
    fib_sample = 34; exp = 35;
test.cpp(60): error: in "test1/_8": check fib_sample == exp has failed [55 != 56]
Failure occurred in a following context:
    fib_sample = 55; exp = 56;

*** 2 failures are detected in the test module "dataset_example68"

Datasets as defined above are constructed before even the test module starts its execution as global objects. This makes impossible to access, from within the dataset generator and during their iteration, elements like argc / argv, the master test suite (and the preprocessed argc / argv), or any other object that has been instanciated after the main of the test module entry.

To overcome this, a delayed dataset instanciation interface has been introduced. This effectively wraps the dataset inside another one, which lazyly instanciates the dataset.

To instanciate a delayed dataset, the boost::unit_test::data::monomorphic::make_delayed function should be used in the BOOST_DATA_TEST_CASE call. The following snippet:

    boost::unit_test::data::make_delayed<custom_dataset>(arg1, ... ), ...)

creates a delayed dataset test case with a generator of type custom_dataset. The generator is lazily constructed with arg1, ....

[Tip] Tip

A detailed example of delayed creation is given in the section about custom command line arguments.

[Tip] Tip

See the class monomorphic::delayed_dataset for more details on the wrapping object.

[2] polymorphic datasets will be considered in the future. Their need is mainly driven by the replacement of the typed parametrized test cases by the dataset-like API.