Re: WIP: hooking parser - Mailing list pgsql-hackers

From Sam Mason
Subject Re: WIP: hooking parser
Date
Msg-id 20090216140922.GV32672@frubble.xen.chris-lamb.co.uk
Whole thread Raw
In response to Re: WIP: hooking parser  (Pavel Stehule <pavel.stehule@gmail.com>)
Responses Re: WIP: hooking parser  (Pavel Stehule <pavel.stehule@gmail.com>)
List pgsql-hackers
On Mon, Feb 16, 2009 at 02:35:54PM +0100, Pavel Stehule wrote:
> attachment contains module that transform every empty string to null.

Why would anyone ever want to do this?  This would appear to break all
sorts of things in very non-obvious ways:
 SELECT CASE s WHEN '' THEN 'empty string' ELSE s END FROM foo; UPDATE foo SET s = NULL WHERE s = '';

would no longer do the expected thing.  It would only do the expected
thing (in my eyes) when strings of zero length were actually being
inserted into the database.  Like:
 INSERT INTO foo (s) VALUES (''); UPDATE foo SET s = '' WHERE s = 'empty string';

Or am I missing something obvious?

--  Sam  http://samason.me.uk/


pgsql-hackers by date:

Previous
From: "Kevin Grittner"
Date:
Subject: Re: SE-PostgreSQL and row level security
Next
From: Pavel Stehule
Date:
Subject: Re: WIP: hooking parser