RE: Question About Space Usage with LONG RAW Column

I think #1 is the cause.  Rows are archived out of the table
periodically, so there have been a lot of deleted rows.  There aren't
that many chained rows.  Thx!

Mark

-----Original Message-----
From: Christian Antognini [mailto:Christian.Antognini@xxxxxxxxxxxx]=20
Sent: Wednesday, October 20, 2004 2:27 PM
To: Mark Strickland
Cc: oracle-l@xxxxxxxxxxxxx
Subject: RE: Question About Space Usage with LONG RAW Column

Hi Mark

In my opinion two possibilities:

1) Many rows have been deleted. Since the delete statements have no
influence on high water mark (which is shown in user_tables.blocks),
empty blocks are not counted.

2) If you have migrated or chained rows (or both at the same time...) by
selecting the rowid you only have the blocks containing the head pieces.


Chris

>-----Original Message-----
>From: oracle-l-bounce@xxxxxxxxxxxxx
[mailto:oracle-l-bounce@xxxxxxxxxxxxx] On Behalf Of Mark
>Strickland
>Sent: 20 October 2004 22:51
>To: oracle-l@xxxxxxxxxxxxx
>Subject: Question About Space Usage with LONG RAW Column
>
>I did a count of the number of blocks of a Kana table (kc_rawtext) with
>this query: =3D20
>
>       select count (distinct substr(rowid,7,9)) from kc_rawtext. =3D20
>
>The result is 2,198,800 blocks.  However, user_tables.blocks shows
>4,113,400 blocks.  I'm trying to figure out the discrepancy.  The table
>has a LONG RAW column and it appears that the column is not the last
>column in the table (column_id for the LONG RAW column is #3 out of
five
>columns).  The table grows quickly.  Can someone help me understand
>where the space is going?  Thx!
>
>Mark Strickland
>--
>http://www.freelists.org/webpage/oracle-l
--
http://www.freelists.org/webpage/oracle-l

Other related posts: