[PostgreSQL-Hackers] Name column

On 24/09/10 13:02, Vlad Arkhipov wrote:
I have just come across a weird thing. It works for any table and seems
to be not documented.

SELECT c.name FROM (VALUES(1, 'A', true)) c;
SELECT c.name FROM pg_class c;

And it does not work in these cases:

SELECT name FROM (VALUES(1, 'A', true));
SELECT name FROM pg_class;

On Fri, Sep 24, 2010 at 6:35 AM, Heikki Linnakangas wrote:
For historical reasons PostgreSQL supports calling a function with a single
argument like "column.function", in addition to "function(column)". There is
a function "name(text)" that casts the input to the 'name' datatype, so your
example casts the row to text and from text to name.

Date: Fri, 24 Sep 2010 08:01:35 -0400
Subject: Re: [HACKERS] Name column
From: robertmhaas@gmail.com
To: heikki.linnakangas@enterprisedb.com
CC: arhipov@dc.baikal.ru; pgsql-hackers@postgresql.org

On Fri, Sep 24, 2010 at 6:35 AM, Heikki Linnakangas
wrote:
I'm starting to wonder if we should think about deprecating this
behavior. It is awfully confusing and unintuitive.

I agree, it is very unintuitive.
+1 for deprecating this behavior.

On Fri, Sep 24, 2010 at 11:43 AM, Tom Lane wrote:
=?iso-8859-1?B?QW5kcukgRmVybmFuZGVz?= <andre.de.camargo.fernandes@hotmail.com> writes:
-1.  There's nothing wrong with the function-as-a-computed-column
feature, and it seems likely that taking it away will break applications.

What we are getting bit by is that I/O coercions to string types can be
specified this way.  Maybe what we ought to do is remove just that one
capability.  It'd be a bit non-orthogonal, but seems fairly unlikely to
break anything, especially since we only began to allow such things
recently (in 8.4 looks like).

Robert Haas writes:
... What evidence do we have that anyone is relying on this
behavior in applications? Every report I've heard of it involved
someone being surprised that it worked that way.

On Fri, Sep 24, 2010 at 11:55 AM, Tom Lane wrote:
Robert Haas <robertmhaas@gmail.com> writes:
So?  There are lots of surprising things in SQL.  And *of course* the
only complaints come from people who didn't know about it, not from
satisfied users.

Robert Haas writes:
I guess that's true, but is this behavior specified in or required by
any SQL standard? Are there other database products that also support
this syntax? Or is this just our own invention?

2010/9/24 Tom Lane <tgl@sss.pgh.pa.us>:
Robert Haas <robertmhaas@gmail.com> writes:
So?  There are lots of surprising things in SQL.  And *of course* the
only complaints come from people who didn't know about it, not from
satisfied users.

The reason people don't know about this feature is that it's so poorly
documented --- there's just one mention buried deep in chapter 35 of
the manual, in a place where most people wouldn't think to look for it.
I'm not quite sure where's a better place though.

regards, tom lane

--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

Pavel Stehule writes:
I dislike this feature too. It is breaking other ANSI SQL feature -
constructors, because it has same syntax tablename(field1, field2,
....).