Re: PL/pgSQL PERFORM with CTE
| От | Pavel Stehule |
|---|---|
| Тема | Re: PL/pgSQL PERFORM with CTE |
| Дата | |
| Msg-id | CAFj8pRATCj1v9VFU3baN8p26A_cpj9=vKEjNMem8Tzkn6LqEbQ@mail.gmail.com обсуждение исходный текст |
| Ответ на | Re: PL/pgSQL PERFORM with CTE ("David E. Wheeler" <david@justatheory.com>) |
| Ответы |
Re: PL/pgSQL PERFORM with CTE
|
| Список | pgsql-hackers |
2013/8/20 David E. Wheeler <david@justatheory.com>
On Aug 20, 2013, at 2:53 PM, Pavel Stehule <pavel.stehule@gmail.com> wrote:I am confused. I do not need to check the result (except via FOUND). But I am sure I can think of other situations where I am calling something where I do not care about the result, even if it returns one.
>> I am passing the values returned from a CTE to a call to pg_notify(). I do not care to collect the output of pg_notify(), which returns VOID.
>
> it is little bit different issue - PL/pgSQL doesn't check if returned type is VOID - it can be allowed, I am thinking. So check of empty result can be enhanced.
When you would to ignore result, then you should to use a PERFORM - actually, it is limited now and should be fixed. Have no problem with it.
I don't would to enable a free unbound statement that returns result.
Regards
Pavel
Best,
David
В списке pgsql-hackers по дате отправления: