Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Strange behaviour

Strange behaviour

From: Carlos Alberto <calberto2312_at_hotmail.com>
Date: 18 Jun 2002 13:48:31 -0700
Message-ID: <72954535.0206181248.c72c06e@posting.google.com>


Hi all,

   I´m experiencing a strange behaviour with Oracle 8.0.6.0.0 : I have a 100 million records table, and based on a condition I loaded 47 million of these records into another table, by INSERT .... SELECT .... command. The big table has 21GB, the new one has 12GB. The structure of them are the same, the only difference is the tablespace where each one resides. The strange is when I run the following script :

set timing on;

select count(1) from BIG_TABLE;

--> 20 minutes

select count(1) from SMALL_TABLE;

--> 1 hour and 10 minutes!!!

   How can this happen? When I run again :

select count(1) from SMALL_TABLE;

--> 6 minutes!!!

   The result now is normal. This behaviour also happens when I run some other SQL, which access some index of the table. The hardware is Sun E10K with EMC Storage. Regarding EMC, the disks, controllers and cache are fine. Can anyone have some hint??

Thanks in advanced,
Carlos Received on Tue Jun 18 2002 - 15:48:31 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US