Re: [HACKERS] Unexpected subselect result.
| От | Vadim B. Mikheev |
|---|---|
| Тема | Re: [HACKERS] Unexpected subselect result. |
| Дата | |
| Msg-id | 34EB98C0.75880BA2@sable.krasnoyarsk.su обсуждение исходный текст |
| Ответ на | Unexpected subselect result. (Keith Parks <emkxp01@mtcc.demon.co.uk>) |
| Список | pgsql-hackers |
Keith Parks wrote:
>
> Hi,
>
> I was having a play with subselects and found this, not very clever,
> nested subselect didn't work as I expected.
>
> Any Ideas?
>
> I also noticed that the subselect regression tests include only
> int and float columns in the tests.
Thanks, Keith!
Vadim
*** nodeSubplan.c.orig Thu Feb 19 08:54:19 1998
--- nodeSubplan.c Thu Feb 19 09:10:25 1998
***************
*** 193,198 ****
--- 193,206 ----
break;
}
+ /*
+ * If this is uncorrelated subquery then its plan will be closed
+ * (see below) and this tuple will be free-ed - bad for not byval
+ * types...
+ */
+ if ( plan->extParam == NULL )
+ tup = heap_copytuple (tup);
+
foreach (lst, node->setParam)
{
ParamExecData *prm = &(plan->state->es_param_exec_vals[lfirsti(lst)]);
В списке pgsql-hackers по дате отправления: