Hi Tom,
> The rationale is "do what the SQL spec says" ;-)
can't argue with the standard ;)
> I seem to recall a previous discussion in the PG lists
Good memory !
Adding 'sql standard' to the search options helped, this issue seems to elude
people every now and then, given (amongst others):
http://archives.postgresql.org/pgsql-hackers/2005-05/msg01457.php
http://archives.postgresql.org/pgsql-general/2006-11/msg00763.php
http://archives.postgresql.org/pgsql-general/2006-11/msg00527.php
One might consider adding a single line to the end of 9.9 of the docs that
warns for this behaviour and/or add the specific example....?
Thanks for your help.
--
Best,
Frank.