On Nov 5, 2010, at 10:49 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Daniel Farina <drfarina@acm.org> writes:
>> Is there any reason why Postgres should not support an "ALTER TABLE
>> tablename [IF EXISTS]" feature?
>
> I think you've phrased the question backwards. Why *should* we support
> that large increment of complexity? The use-cases seem pretty few and
> far between.
Obviously we have different definitions of "a large increment in complexity".
...Robert