RE: Parallel INSERT (INTO ... SELECT ...)

Hi,

When developing the reloption patch, I noticed some issues in the patch.

1).
> - Reduce Insert parallel-safety checks required for some SQL, by noting
> that the subquery must operate on a relation (check for RTE_RELATION in
> subquery range-table)

+ foreach(lcSub, rte->subquery->rtable)
+ {
+ rteSub = lfirst_node(RangeTblEntry, lcSub);
+ if (rteSub->rtekind == RTE_RELATION)
+ {
+ hasSubQueryOnRelation = true;
+ break;
+ }
+ }
It seems we can not only search RTE_RELATION in rtable,
because RTE_RELATION may exist in other place like:

---
--** explain insert into target select (select * from test);
Subplan's subplan

--** with cte as (select * from test) insert into target select * from cte;
In query's ctelist.
---

May be we should use a walker function [1] to
search the subquery and ctelist.

2).

+--
+-- Test INSERT into temporary table with underlying query.
+-- (should not use a parallel plan)
+--

May be the comment here need some change since
we currently support parallel plan for temp table.

3)
Do you think we can add a testcase for foreign-table ?
To test parallel query with serial insert on foreign table.

[1]
static bool
relation_walker(Node *node)
{
if (node == NULL)
return false;

else if (IsA(node, RangeTblEntry))
{
RangeTblEntry *rte = (RangeTblEntry *) node;
if (rte->rtekind == RTE_RELATION)
return true;

return false;
}

else if (IsA(node, Query))
{
Query *query = (Query *) node;

/* Recurse into subselects */
return query_tree_walker(query, relation_walker,
NULL, QTW_EXAMINE_RTES_BEFORE);
}

/* Recurse to check arguments */
return expression_tree_walker(node,
relation_walker,
NULL);
}

Best regards,
houzj