=?iso-8859-1?Q?Rolf_L=FCttecke?= <rolf.luettecke@michael-telecom.de> writes:
>> I can't find out why the following command give different results:
> ALTER TABLE ADD COLUMN doesn't support adding a default. (There is a
> test in the code that's supposed to kick out an error message telling
> you so, but it doesn't seem to be working. Need to look at that.)
I've committed fixes for that problem in 7.2.
> Hmm ... I wonder why ADD COLUMN doesn't just call the SET DEFAULT
> routine to handle adding the default ...
On looking at the archives (see pgsql-hackers around 17-Jul-01),
I see that this wasn't done because it's not an SQL92-compliant
behavior. Our current implementation of ADD COLUMN can't support
setting the new column's entries to anything except NULL, but the
spec says that if there's a DEFAULT then the column should initially
be filled with that value.
regards, tom lane
