Re: DROP COLUMN

On Wed, 2002-07-17 at 08:26, Tom Lane wrote:
> Hannu Krosing <hannu(at)tm(dot)ee> writes:
> > Also, as we have nothing like Oracles ROWNR, I think it will be quite
> > hard to have colnums without gaps in the system views, so we could
> > perhaps have a stopgap solution of adding logical column numbers (
> > (pg_attribute.attlognum) that will be changed every time a col is
> > added/dropped just for that purpose.
>
> [ thinks... ] I don't believe this would make life any easier, really.
> Inside the backend it's not much help, because we still have to look
> at every single attnum reference to see if it should be logical or
> physical attnum.

I meant this as a workaround for missing ROWNR pseudocolumn.

All backend functions would still use real attnum's. And I doubt that
backend will ever work though system views.

Adding them should touch _only_ CREATE TABLE, ADD COLUMN, DROP COLUMN
plus the system views and possibly output from SELECT(*), if we allow
logical reordering of columns by changing attlognum.

Of course we would not need them if we had ROWNR (or was it ROWNUM ;),
except for the hypothetical column reordering (which would be useful for
ALTER COLUMN CHANGE TYPE too)

> On the client side it seems promising at first sight
> ... but the client will still break if it tries to correlate the
> logical colnum it sees with physical colnums in pg_attrdef and other
> system catalogs.

One can alway look it up in pg_attribute ;)

Just remember to use attlognum _only_ for presentation.

> Bottom line AFAICT is that it's a lot of work and a lot of code
> to examine either way :-(

Yes, I see that it can open another can of worms .

--------------
Hannu