On Fri, Oct 27, 2023 at 10:41 PM Tom Lane <tgl@sss.pgh.pa.us> wrote:
Richard Guo <guofenglinux@gmail.com> writes: > Alternatively, can we look at subroot->parse->targetList instead of > subquery->targetList where we call estimate_num_groups on the output of > the subquery?
Please read the comment just above that.
Hmm, I understand that we cannot look at subroot->processed_tlist here because of the "varno 0" hack explained in the comment. But what I proposed is to look at subroot->parse->targetList, which should be fine because the tlist generated by generate_append_tlist would only be returned into processed_tlist not the original parse->targetList. For instance, the query below still works if we look at subroot->parse->targetList, even though the subquery is itself a setop.
(SELECT 1,2,3 UNION SELECT 4,5,6 ORDER BY 1,2) INTERSECT SELECT 4,5,6; ?column? | ?column? | ?column? ----------+----------+---------- 4 | 5 | 6 (1 row)
