Re: Oracle 9.2.0.1.0 For Windows performance question

From: Sybrand Bakker <gooiditweg_at_sybrandb.demon.nl>
Date: Sat, 23 Aug 2003 20:11:36 +0200
Message-ID: <debfkv4b5sb32gvlgr6tvhdm6me2ne4qur@4ax.com>

On 23 Aug 2003 09:11:39 -0700, grzybek_at_mikrus.pw.edu.pl (Aguyngueran) wrote:

>> SELECT
>> kod_w || kod_p || kod_g || rodzaj_g, sum(decode(e,'N',1,0)) N,
>> sum(decode(e,'U',1,0) U -- etc
>> from tabelka
>> group by kod_w ||kod_p||kod_g||rodzaj_g
>>
>> and you can throw away the trash you are using now.
>>
>Thank You very much. I didn't know about DECODE function.
>But what about SELECT count(*) from TABELKA; It takes quite long time

The duration of a count(*) is dependent of the size of the table and the location of High Water Mark. If you have a table of 1000 blocks, and 0 records, but the High Water Mark is at block 995, all 995 blocks will be read.
Also your db_file_multiblock_read_count may be inadequate per se, or be inadequate compared to the size of a single extent of your table

Sybrand Bakker, Senior Oracle DBA

To reply remove -verwijderdit from my e-mail address Received on Sat Aug 23 2003 - 13:11:36 CDT