Re: Make foo=null a warning by default.

Поиск
Список
Период
Сортировка
От Heikki Linnakangas
Тема Re: Make foo=null a warning by default.
Дата
Msg-id 965f0642-09e2-9729-9aae-e0d088ea4598@iki.fi
обсуждение исходный текст
Ответ на Make foo=null a warning by default.  (David Fetter <david@fetter.org>)
Ответы Re: Make foo=null a warning by default.  (Fabien COELHO <coelho@cri.ensmp.fr>)
Re: Make foo=null a warning by default.  (Robert Haas <robertmhaas@gmail.com>)
Re: Make foo=null a warning by default.  (Tom Lane <tgl@sss.pgh.pa.us>)
Список pgsql-hackers
On 16/07/18 04:40, David Fetter wrote:
> Folks,
> 
> Per a discussion with Andrew Gierth and Vik Fearing, both of whom
> helped make this happen, please find attached a patch which makes it
> possible to get SQL standard behavior for "= NULL", which is an error.
> It's been upgraded to a warning, and can still be downgraded to
> silence (off) and MS-SQL-compatible behavior (on).

I don't agree with changing the default to 'warn'. "foo = NULL" is 
perfectly legal SQL, even if it's not very useful in practice.

The use case for transform_null_equals='warn' that I see is to wean off 
an application that uses transform_null_equals='on', to find all the 
queries that rely on it. But I'm not too excited about that either. The 
documented case for using transform_null_equals='on' is compatibility 
with Microsoft Access, and this wouldn't help with that. Besides, it's 
easy enough to grep the application or the server log for " = NULL", to 
find all such queries.

- Heikki


В списке pgsql-hackers по дате отправления:

Предыдущее
От: Paul Muntyanu
Дата:
Сообщение: Parallel queries in single transaction
Следующее
От: Fabien COELHO
Дата:
Сообщение: Re: Make foo=null a warning by default.