check_correlations_in_subquery should not check uncorrelated plan nodes #14461

#13523 fixed some queries with unnest in subquery without adding support for unnest decorelation (which is possible; tracked by #11773) by changing the check function not to fail on unnest nodes.
It would be better, however, if the check function didn't check uncorrelated plan nodes at all. An uncorrelated part of the plan is a unit that doesn't need to be analyzed.

relates to

• [dbolympics] General framework to decorrelate the subqueries #5492
• Support correlated UNNEST in subquery #11773