If I were you I'd try to analyze both tables (prod and test), posibly on
a small sample, to check whether the average row length is identical.
The 'comment' field, which can be close enough to 60%, may be NULL in
your test database and full in most cases in prod. Something else one
could think of is a PCTFREE difference. It may be higher on your prod
database. There may also be differences in the per-tablespace minimum
allocation values.

S Faroult

"Jesse, Rich" wrote:
> 
> Hi all,
> 
> 8.1.7.2.0 on HP/UX 11.0 -- A developer tested a table recreate in our test
> DB.  Here's the scenario:
> 
> CREATE TABLE MY_BIG_TABLE (
>   FISCAL_YEAR   NUMBER (5),
>   PERIOD        CHAR (2),
>   ACCOUNTNO     CHAR (12),
>   TRANSTYPE     CHAR (2),
>   TRANSQTY      FLOAT,
>   TRANSAMNT     FLOAT,
>   COMMENT_TEXT  CHAR (30),
>   TRANSDATE     DATE)
>    TABLESPACE QT_APPS1
>    PCTFREE 10
>    PCTUSED 40
>    INITRANS 2
>    MAXTRANS 255
>   STORAGE (
>    INITIAL 209715200
>    NEXT 10485760
>    PCTINCREASE 0
>    MINEXTENTS 1
>    MAXEXTENTS 249
>    FREELISTS 1 FREELIST GROUPS 1 )
>    NOCACHE;
> 
> The table previously had a total of 14 extents, giving it a size of 330MB.
> The dev TRUNCATED the table (I don't know if "REUSE STORAGE" was used), and
> added 4,054,632 (4M) rows.  The table is still at 14 extents and 330MB.
> 
> So, all's well in test, the dev did the same in production.  The row count
> was *slightly* higher, at 4.069,106, but the table size jumped 60% to 560MB.
> The DB_BLOCK_SIZE on both DBs is 8K.  Of course, this filled up the TBS and
> caused havoc.
> 
> I looked in DBA_AUDIT_TRAIL and DBA_TAB_MODIFICATIONS, but I can't find any
> significant difference between what was done to the table in test and prod.
> The table in production has NO deletes recorded in DBA_TAB_MODIFICATIONS,
> just inserts.
> 
> Can anyone think of a scenario as to why this table would grow in prod but
> not test with relatively the same number of rows and the exact same table
> layout???  The only thing I can think of is that a "REUSE STORAGE" was
> issued on one TRUNCATE, but not another, but I still don't see how that
> could account for the table growth.
> 
> I'm going thru LogMiner now, but as our test DB is in archivelog mode, I can
> only look in production (and it's taking forever!).
> 
> TIA!
> 
>
-- 
Please see the official ORACLE-L FAQ: http://www.orafaq.com
-- 
Author: Stephane Faroult
  INET: [EMAIL PROTECTED]

Fat City Network Services    -- (858) 538-5051  FAX: (858) 538-5051
San Diego, California        -- Public Internet access / Mailing Lists
--------------------------------------------------------------------
To REMOVE yourself from this mailing list, send an E-Mail message
to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB ORACLE-L
(or the name of mailing list you want to be removed from).  You may
also send the HELP command for other information (like subscribing).

Reply via email to