Re: Document NULL - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Document NULL
Date
Msg-id 1649111.1741725829@sss.pgh.pa.us
Whole thread Raw
In response to Re: Document NULL  ("David G. Johnston" <david.g.johnston@gmail.com>)
Responses Re: Document NULL
List pgsql-hackers
"David G. Johnston" <david.g.johnston@gmail.com> writes:
> On Tue, Mar 11, 2025 at 1:19 PM Marcos Pegoraro <marcos@f10.com.br> wrote:
>> I understand that this page we have more NULL values than everywhere else
>> combined, but it's a good opportunity to set a standard for all
>> documentation

> If someone writes a patch to make everywhere <null> and changes this page
> in the process I wouldn't complain.  For now there is no such standard, we
> use empty string everywhere else, and that isn't a good choice here IMO.
> It is probably a good choice for elsewhere since the empty string is our
> default.  While specific to this page \N reads the best for these examples
> IMO.  But I'll concede to whatever if others think this choice is a
> review/commit blocker.

I think that idea (changing all the docs) is a complete nonstarter
because people would not understand why the results they get don't
look like what it says in the docs.  Of course, we could fix that
by changing the factory default for \pset null, but that seems like
even more of a nonstarter.

We can probably get away with having one page that assumes a
different \pset null setting, as long as there's a clear <note>
about it at the top of the page.

            regards, tom lane



pgsql-hackers by date:

Previous
From: "David G. Johnston"
Date:
Subject: Re: Document NULL
Next
From: "Devulapalli, Raghuveer"
Date:
Subject: RE: Improve CRC32C performance on SSE4.2