From: | Alex Pilosov <alex(at)pilosoft(dot)com> |
---|---|
To: | mlw <markw(at)mohawksoft(dot)com> |
Cc: | pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: Changing data types |
Date: | 2001-09-25 01:41:44 |
Message-ID: | Pine.BSO.4.10.10109242133390.28103-100000@spider.pilosoft.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgadmin-hackers pgsql-hackers |
On Mon, 24 Sep 2001, mlw wrote:
> To be honest I am very surprised that MS SQL supports that, but then
> again Microsoft is so used to doing everything so utterly wrong, they
> have to design all their products with the ability to support
> fundamental design error corrections on the fly.
>
> I would be surprised if Oracle, DB2, or other "industrial grade"
> databases could do this. Needing to change a column from a varchar to
> an integer is a huge change and a major error in design.
Actually they do. Its not a such a great deal, same as adding a column and
dropping a column. If you can do that, you can do modification of type.
The sticky thing is dropping a column. There are two options, and
postgresql developers just can't make up their mind :P)
a) keep old column data in database (wasted space, but fast)
b) immediately 'compress' table, removing old data (slow, needs a lot of
space for compression)
Option a) was implemented once, but kludgy, and had a few kinks, and it
was removed. Option b) plain sucks :P)
-alex
From | Date | Subject | |
---|---|---|---|
Next Message | Rod Taylor | 2001-09-25 02:22:09 | Re: Changing data types |
Previous Message | mlw | 2001-09-25 01:25:10 | Re: Changing data types |
From | Date | Subject | |
---|---|---|---|
Next Message | Hiroshi Inoue | 2001-09-25 01:42:04 | Re: ODBC driver flakieness |
Previous Message | mlw | 2001-09-25 01:25:10 | Re: Changing data types |