On Wednesday, August 7, 2024, dfgpostgres <dfgpostgr...@gmail.com> wrote:
> > (select > domain_name, > sum(total_tests) as total_tests, > sum(tests_completed) as tests_completed, > sum(tests_passed) as tests_passed, > sum(tests_failed) as tests_failed, > (select count(*) from dispatch_tracker where > regression_name=rt.regression_name and domain_name=rt.domain_name and > dispatch_status='Y') as dispatched > from > regr.dispatch_tracker rt where rt.regression_name='2024_08_ > 02_10_32_53_soundwave__er_common_regression__CL2017473_z1_soundwave_adm' > group by rollup(rt.domain_name) order by rt.domain_name ASC NULLS LAST) > d; > Either add regression_name to the group by as the error hints at you, or since you are already grouping implicitly by that (by virtue of the where clause filter) and domain_name just count the number of dispatch_status=Y in the group: count(*) filter (where dispatch_status = ‘Y’) The option that avoids the subquery is arguably better. Though I’d probably still include the regression_name in the output anyway - why hide what you are filtering on. David J.
