Re: BUG #14268: NULL parameter conversion

2016-07-28 23:36 GMT+03:00 David G. Johnston <david(dot)g(dot)johnston(at)gmail(dot)com>:

> ​Please don't top-post.
>
> On Thu, Jul 28, 2016 at 4:16 PM, Jordan Gigov <coladict(at)gmail(dot)com> wrote:
>
>> I don't see how that helps your argument that NULLs shouldn't be
>> convertible.
>>
>
> ​Your previous response didn't help your argument that they should be
> convertible.​
>
> ​NULLs aren't convertible - and while you can argue for a better policy
> this one isn't wrong - I don't actually need to make arguments supporting
> that fact. The underlying point, though, is the ability for types to be
> converted between one another is a property of the type itself and not any
> specific value that type may take on.
>
> In any case - if you wish to turn this into a discussion I'd recommend
> putting together a more detailed argument and posting in on
> pgsql-general(at)postgresql(dot)org
>
> David J.
> ​
>
>

While I don't plan on spending 180€ for the active ISO 9075 specifications,
I did find a working draft of 9075-2 where under section 6.13 <cast
specification> in the general rules it says:
"If the <cast operand> specifies NULL, then the result of CS is the null
value and no further General Rules of this Subclause are applied."
You asked for an authority. This is the closest I can get, without
unnecessary (at least for me, as someone who isn't implementing an SQL
database) expenses.
I got it from http://www.wiscorp.com/SQLStandards.html the link that says
"SQL:20nn Working Draft Documents". It doesn't have one of those red
markers that I assume are for proposed and not finalized changes.