an already existing alter table drop column ?!?!?!

On Fri, 21 Sep 2001, Gowey, Geoffrey wrote:

While looking through the code I found an already existing alter table drop
column in src/backend/parser/gram.y. However, when I try to run it in psql
it comes back with a not implemented. Was the left hand not talking to the
right hand when this was coded or is there more to this?

IIRC, it was not enabled pending further discussion of the behavior.

Stephan Szabo wrote:

On Fri, 21 Sep 2001, Gowey, Geoffrey wrote:

While looking through the code I found an already existing alter table drop
column in src/backend/parser/gram.y. However, when I try to run it in psql
it comes back with a not implemented. Was the left hand not talking to the
right hand when this was coded or is there more to this?

IIRC, it was not enabled pending further discussion of the behavior.

As to 'DROP COLUMN', I neglected to remove _DROP_COLUMN_HACK__
stuff which has no meaing now sorry. I would remove it after
the 7.2 release.

regards,
Hiroshi Inoue

Does this mean that the code is or isn't usable?

Geoff

-----Original Message-----
From: Hiroshi Inoue [mailto:Inoue@tpf.co.jp]
Sent: Sunday, September 23, 2001 11:02 PM
To: Stephan Szabo
Cc: Gowey, Geoffrey; pgsql-hackers@postgresql.org
Subject: Re: [HACKERS] an already existing alter table drop column ?!?!?!

Stephan Szabo wrote:

On Fri, 21 Sep 2001, Gowey, Geoffrey wrote:

While looking through the code I found an already existing alter table

drop

column in src/backend/parser/gram.y. However, when I try to run it in

psql

it comes back with a not implemented. Was the left hand not talking to

the

right hand when this was coded or is there more to this?

IIRC, it was not enabled pending further discussion of the behavior.

As to 'DROP COLUMN', I neglected to remove _DROP_COLUMN_HACK__
stuff which has no meaing now sorry. I would remove it after
the 7.2 release.

regards,
Hiroshi Inoue

isn't

On Mon, 24 Sep 2001, Gowey, Geoffrey wrote:

damn. Is there a usable version being worked around? If so, how long before
it is available?

Geoff

-----Original Message-----
From: Marc G. Fournier [mailto:scrappy@hub.org]
Sent: Monday, September 24, 2001 1:14 PM
To: Gowey, Geoffrey
Cc: 'Hiroshi Inoue'; Stephan Szabo; pgsql-hackers@postgresql.org
Subject: Re: [HACKERS] an already existing alter table drop column
?!?!?!

isn't

On Mon, 24 Sep 2001, Gowey, Geoffrey wrote:

damn. Is there a usable version being worked around? If so, how long before
it is available?

too many dissending opinions on how it should (and shouldn't) be done ...
the one that was put in pre-maturely needed 2x the disk space to do ... so
if you had a multi-gig table, you had to have enough free space to store a
second copy in order to remove a column ...

there was a bunch of us that felt that it could be done better using a
'flag' on the column so that it was hidden ... and some that thought ...
etc, etc ...

I do like the flag idea as a temporary patch until a fully workable solution
is available. Just run a delete on the column and then flag it so it takes
up no space and is hidden. By doing this now a more perfect solution can be
found later.

Geoff

-----Original Message-----
From: Marc G. Fournier [mailto:scrappy@hub.org]
Sent: Monday, September 24, 2001 1:28 PM
To: Gowey, Geoffrey
Cc: 'Hiroshi Inoue'; Stephan Szabo; pgsql-hackers@postgresql.org
Subject: RE: [HACKERS] an already existing alter table drop column
?!?!?!

On Mon, 24 Sep 2001, Gowey, Geoffrey wrote:

damn. Is there a usable version being worked around? If so, how long

before

it is available?

too many dissending opinions on how it should (and shouldn't) be done ...
the one that was put in pre-maturely needed 2x the disk space to do ... so
if you had a multi-gig table, you had to have enough free space to store a
second copy in order to remove a column ...

there was a bunch of us that felt that it could be done better using a
'flag' on the column so that it was hidden ... and some that thought ...
etc, etc ...

Also, rename the column to something else so a new one can be created in
it's place with the same name. Do columns have their own oid's? If so this
plus some random chars could be the new name.

Geoff

-----Original Message-----
From: Gowey, Geoffrey
Sent: Monday, September 24, 2001 1:35 PM
To: 'Marc G. Fournier'; Gowey, Geoffrey
Cc: 'Hiroshi Inoue'; Stephan Szabo; pgsql-hackers@postgresql.org
Subject: RE: [HACKERS] an already existing alter table drop column
?!?!?!

I do like the flag idea as a temporary patch until a fully workable solution
is available. Just run a delete on the column and then flag it so it takes
up no space and is hidden. By doing this now a more perfect solution can be
found later.

Geoff

-----Original Message-----
From: Marc G. Fournier [mailto:scrappy@hub.org]
Sent: Monday, September 24, 2001 1:28 PM
To: Gowey, Geoffrey
Cc: 'Hiroshi Inoue'; Stephan Szabo; pgsql-hackers@postgresql.org
Subject: RE: [HACKERS] an already existing alter table drop column
?!?!?!

On Mon, 24 Sep 2001, Gowey, Geoffrey wrote:

damn. Is there a usable version being worked around? If so, how long

before

it is available?

too many dissending opinions on how it should (and shouldn't) be done ...
the one that was put in pre-maturely needed 2x the disk space to do ... so
if you had a multi-gig table, you had to have enough free space to store a
second copy in order to remove a column ...

there was a bunch of us that felt that it could be done better using a
'flag' on the column so that it was hidden ... and some that thought ...
etc, etc ...