Re: quick question on decompressing a table...

  • From: Tanel Põder <tanel.poder.003@xxxxxxx>
  • To: <oracle-l@xxxxxxxxxxxxx>
  • Date: Thu, 19 Feb 2004 00:55:39 +0200

If you want to decompress the existing blocks as well, then you can just use
"alter table move nocompress", with nologging and parallel if you want. (but
test it out on a test table first, I had some problems with it in 9.2.0.1).

Tanel.

----- Original Message ----- 
From: "Freeman Robert - IL" <FREEMANR@xxxxxxxx>
To: "'Chris Stephens '" <ChrisStephens@xxxxxxxxxx>; <oracle-l@xxxxxxxxxxxxx>
Sent: Thursday, February 19, 2004 12:38 AM
Subject: RE: quick question on decompressing a table...


> Someone might disagree with me, but it strikes me that this is *much* more
> work than it's worth. Simply alter the table to not use compression. New
> data will not be compressed and While the data thats in it right now might
> be in compressed format, any change to that data will result in it's
> permenant uncompression. Since there isn't much data in it at this point,
I
> would not think that this would be that big of a deal....
>
> RF
>
> -----Original Message-----
> From: Chris Stephens
> To: oracle-l@xxxxxxxxxxxxx
> Sent: 2/18/2004 4:33 PM
> Subject: quick question on decompressing a table...
>
> We have a fact table that was mistakenly created with the compression
> option.
> Fortunately there is not (too) much data in the table as of yet.
>
> The table is partitioned and have bitmap indexes on all the foreign
> keys.
>
> Downtime isn't a huge deal...but minimizing it would be a positive
> thing.
>
>
>
> The easiest way I can think of to correct this is the following:
>
>
>
> Create table temp as select * from comp_table nologging;
>
>
>
> Drop table comp_table;
>
>
>
> Recreate table comp_table
>
>
>
> Insert /*+ append */ into comp_table select * from temp_table;
>
>
>
> Rebuild bitmaps
>
>
>
> Regrant privileges.
>
>
>
>
>
> ...any faster/easier way to do this?   ...any gotchas working with
> compressed tables (not sure how it would relate to this situation)?
>
>  ...I'm no expert on table compression and as it stands now, r'ing tfm
> at
> this point isn't won't be worth any time savings.  I'm mostly just (a
> little) curious.
>
>
>
> Thanks for any suggestions
>
> Chris
>
>
>
> ----------------------------------------------------------------
> Please see the official ORACLE-L FAQ: http://www.orafaq.com
> ----------------------------------------------------------------
> To unsubscribe send email to:  oracle-l-request@xxxxxxxxxxxxx
> put 'unsubscribe' in the subject line.
> --
> Archives are at //www.freelists.org/archives/oracle-l/
> FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
> -----------------------------------------------------------------
> ----------------------------------------------------------------
> Please see the official ORACLE-L FAQ: http://www.orafaq.com
> ----------------------------------------------------------------
> To unsubscribe send email to:  oracle-l-request@xxxxxxxxxxxxx
> put 'unsubscribe' in the subject line.
> --
> Archives are at //www.freelists.org/archives/oracle-l/
> FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
> -----------------------------------------------------------------
>


----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to:  oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
--
Archives are at //www.freelists.org/archives/oracle-l/
FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------

Other related posts: