Bad scaling of data-driven tests

A data-driven test covering an xrange of 1 million numbers generates 1 million xUnit XML reports.
This brings most tooling to a crawl, in particular Jenkins.

Each data point should not generate a new test case.
The documentation suggests that this should scale well and that it is valid to make tests like this.
https://www.boost.org/doc/libs/1_69_0/libs/test/doc/html/boost_test/tests_organization/test_cases/test_case_generation.html