Re: postgres_fdw: Oddity in pushing down inherited UPDATE/DELETE joins to remote servers

On Fri, May 11, 2018 at 8:46 AM, Etsuro Fujita
<fujita(dot)etsuro(at)lab(dot)ntt(dot)co(dot)jp> wrote:
> I added an assertion test to make_modifytable to match that in
> create_modifytable_path. Attached is an updated version of the patch.

Committed. Was it just good luck that this ever worked at all? I mean:

- if (rti < root->simple_rel_array_size &&
- root->simple_rel_array[rti] != NULL)
+ if (rti < subroot->simple_rel_array_size &&
+ subroot->simple_rel_array[rti] != NULL)
{
- RelOptInfo *resultRel = root->simple_rel_array[rti];
+ RelOptInfo *resultRel = subroot->simple_rel_array[rti];

fdwroutine = resultRel->fdwroutine;
}
else
{
- RangeTblEntry *rte = planner_rt_fetch(rti, root);
+ RangeTblEntry *rte = planner_rt_fetch(rti, subroot);

...an RTI is only meaningful relative to a particular PlannerInfo's
range table. It can't be right to just take an RTI for one
PlannerInfo and look it up in some other PlannerInfo's range table.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company