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 -> Newbie: indexing long raw table

Newbie: indexing long raw table

From: DSmith <dsmith2005_at_yahoo.com>
Date: Wed, 27 Jan 1999 17:49:38 -0500
Message-ID: <36b0422a.0@208.206.112.5>


All,

I have a long raw table on an O 8.0.4 DB, that is used to store various images, it is stored in the "data" tablespace. There is a primary key index on the "img_name" column that is stored in "indx" tablespace. The application code performs selects on the table based on the primary key of "img_name".

We have begun to load several thousand records at ~ 50K bytes each and I've noticed that whenever I do a count(*) OR count(name) it appears to invoke a full-table scan. Is this normal? If not, what can I do to fix it?

It seems that only the index should be scanned to determine the record count, particularly when using the indexed column "img_name".

I've also noticed that SYSTEM tablespace has grown dramatically (as reported by Storage Manager) to a high-water mark of over 3GB during the loading process. Any thoughts (the only users that have SYSTEM as their default tablespace are SYS, SYSTEM, and SNMP)?

Any help would be greatly appreciated.

--David Smith
dsmith_at_hpti.com Received on Wed Jan 27 1999 - 16:49:38 CST

Original text of this message

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