While grouping it also consider tests which are skipped & result of that is uneven groups

[rajeev-ranjan7]

Since grouping does not factor in that there will be some tests which will get skipped and due to that when grouping happened, one group will have 2 tests run and others will have 10 tests run. So overall test execution time gets impacted and maybe more with more number of groups.

[yoram-s]

I think the solution should be to add the following, so the collection hook will be executed after the filtering was applied

@pytest.hookimpl(trylast=True)
def pytest_collection_modifyitems(session, config, items):

[JulienGout]

+1 , since this is a real show stopper for using this plugin in our project.