=?iso-8859-2?Q?Wojciech_Strza=B3ka?= <[EMAIL PROTECTED]> writes:
>> To make that happen would require (at least) a full table scan. I think
>> most people are more interested in DROP COLUMN being a cheap operation
>> than in having the space be reclaimed quickly.
>> For a comparison point: large
> To make that happen would require (at least) a full table scan. I think
> most people are more interested in DROP COLUMN being a cheap operation
> than in having the space be reclaimed quickly.
> For a comparison point: large field values that don't happen to get
> toasted don't vanish immedia
=?iso-8859-2?Q?Wojciech_Strza=B3ka?= <[EMAIL PROTECTED]> writes:
> In my opinion the fact that dropping column doesn't release it's toastable
> resources is a bug.
To make that happen would require (at least) a full table scan. I think
most people are more interested in DROP COLUMN being a cheap
I've found some strange behavoiur of TOAST'able tables.
1. Lets create table with toastable column
CREATE table toastable (
x int ,
y text
);
2. Check toast size - as the table is empty it's size 0 - OK
SELECT relname, pg_relation_size(oid) FROM pg_class where oid=(select
reltoastrelid from
