Re: BUG #19037: Planner fails on estimating array length with "no relation entry" error

Richard Guo <guofenglinux@gmail.com> writes:
> Here is the patch that passes root as NULL to cost_qual_eval() in
> cost_subplan(), along with some commentary about why.

v2 LGTM.  I did think of a way to make the test case a little
less busy-looking:

select (1 = any(array_agg(f1))) = any (select false) from int4_tbl;

but that's just cosmetic.

            regards, tom lane