Tom Lane wrote:
>
>> ALTER ADD COLUMN doesn't touch any tuples, and you're right that it's
> critically dependent on heap_getattr returning NULL when an attribute
> beyond the number of attributes actually present in a tuple is accessed.
> That's a fragile and unclean implementation IMHO --- see past traffic
> on this list.
Short of redesigning the whole storage format I can see no better way to
allow
ALTER ADD COLUMN in any reasonable time. And I cna see no place where
this is
more "fragile and unclean implementation" than any other in postgres --
OTOH it is quite hard for me to "see the past traffic on this list" as
my
"PgSQL HACKERS" mail folder is too big for anything else then grep ;)
The notion that anything not stored is NULL seems so natural to me that
it
is very hard to find any substantial flaw or fragility with it.
--------------
Hannu