> -----Original Message-----
> From: Stephan Szabo [mailto:sszabo@megazone.bigpanda.com]
> Sent: Wednesday, October 19, 2005 12:39 PM
> To: Dann Corbit
> Cc: Marc G. Fournier; Richard_D_Levine@raytheon.com; pgsql-
> general@postgresql.org
> Subject: Re: [pgsql-advocacy] [GENERAL] Oracle buys Innobase
>
> On Wed, 19 Oct 2005, Dann Corbit wrote:
>
> > Yes, clearly that is the wrong result according to the SQL standard.
> >
> > Here is a SQL*Server query:
> > select 1 where 'a' = 'a ' AND 'a' = 'a ' AND 'a ' = 'a '
> >
> > It returns (correctly): 1
>
> Doesn't that depend on the collating sequence in use, or is a NO PAD
> collating sequence not allowed here?
If the implementation defines constants as NO PAD and the implementation
defined pad character is something other than space, then they could
compare unequal.
I would find that implementation disturbing. But I am easily bent out
of shape.
The attached HTML file in my earlier post is the official quote from the
SQL 99 standard. That is the formal and correct definition, far
superior to my off the cuff approximations.