Re: jsonb_array_elements issue

Поиск
Список
Период
Сортировка
От Mephysto
Тема Re: jsonb_array_elements issue
Дата
Msg-id CAG0sfBXGd3CZj9-OnyuEtG-kqS_XBvOFZ1U8CkZT71GpoCvoog@mail.gmail.com
обсуждение исходный текст
Ответ на Re: jsonb_array_elements issue  (Tom Lane <tgl@sss.pgh.pa.us>)
Список pgsql-bugs
Thank you very much for your explanation Tom.

Very kind regards.

Meph

On 16 Aug 2016 5:56 p.m., "Tom Lane" <tgl@sss.pgh.pa.us> wrote:

> Mephysto <mephystoonhell@gmail.com> writes:
> > In my previous email I forgot to say that I create a cast in my database
> as
> > I made with JSON. The cast is this:
>
> > CREATE CAST (text AS JSONB) WITHOUT FUNCTION AS IMPLICIT;
>
> Well, that's rather a critical bit of information :-(
>
> > If you create this cast and you launch my select without the explicit
> cast,
> > you should get the error. With this cast created I reported some other
> > issues, so at this point my question is: can I use this implicit cast or
> is
> > better to explicit it anytime is needed?
>
> This cast definition is completely broken; it's astonishing that you
> haven't had outright crashes, because text and jsonb do NOT have the
> same underlying representation, which is what would be required to
> make a cast WITHOUT FUNCTION work correctly.  But this certainly
> explains 'unknown type of jsonb container' errors.
>
> You could fix that by providing a suitable casting function.  I'm still
> dubious that making it an implicit cast is a good idea though.  The
> trouble with implicit casts is that they tend to kick in when you were
> not expecting them to.  Past experience has suggested a rule of thumb
> that implicit cross-type-category casts are best avoided, and I'd
> certainly call this a type category crossing.
>
>                         regards, tom lane
>

В списке pgsql-bugs по дате отправления:

Предыдущее
От: Tom Lane
Дата:
Сообщение: Re: jsonb_array_elements issue
Следующее
От: Andres Freund
Дата:
Сообщение: Re: BUG #14243: pg_basebackup failes by a STATUS_DELETE_PENDING file