On Fri, Oct 14, 2022 at 5:24 PM Alvaro Herrera <[email protected]>
wrote:
> Actually, I hadn't realized that the originally submitted patch had the
> test in postgres_fdw only, but we really want it to catch any FDW, so it
> needs to be somewhere more general. The best place I found to put this
> test is in make_modifytable ... I searched for some earlier place in the
> planner to do it, but couldn't find anything.
>
> So what do people think about this?
Good point. I agree that the test should be in a more general place.
I wonder if we can make it earlier in grouping_planner() just before we
add ModifyTablePath.
--- a/src/backend/optimizer/plan/planner.c
+++ b/src/backend/optimizer/plan/planner.c
@@ -1772,6 +1772,17 @@ grouping_planner(PlannerInfo *root, double
tuple_fraction)
/* Build per-target-rel lists needed by ModifyTable */
resultRelations = lappend_int(resultRelations,
resultRelation);
+ if (parse->commandType == CMD_MERGE &&
+ this_result_rel->fdwroutine != NULL)
+ {
+ RangeTblEntry *rte = root->simple_rte_array[resultRelation];
+
+ ereport(ERROR,
+ errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
+ errmsg("cannot execute MERGE on relation \"%s\"",
+ get_rel_name(rte->relid)),
+ errdetail_relkind_not_supported(rte->relkind));
+ }
Thanks
Richard