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 -> Re: Count(*) is very slow

Re: Count(*) is very slow

From: Cheng-Jih Chen <cjc_at_interport.net>
Date: 22 Oct 1998 13:27:00 -0400
Message-ID: <70npt4$lcm$1@interport.net>


In article <362f1a54.19073887_at_news.siol.net>, Jurij Modic <jmodic_at_src.si> wrote:
>
>>Andreas Kyritz wrote in message <70mv73$p34$1_at_news.ipf.net>...
>>
>>>I've got a mail with the right idea from Dave Smith from Australia.
>>>The trick to do 'SELECT COUNT(*) FROM ATABLE' fast is, to include:
>>> fast_full_scan_enabled=true
>>>to the init file.
>>>That's all and now it takes less than a second. 8-))
>
>This particular init parameter is realy new to Oracle8, but "fast
>full-index scan" is available with Oracle 7.3.3 and above. There is
>even a new hint "INDEX_FFS (table, index)" to force a fast full-index
>scan. There was a very good article by Guy Harrison about this new
>future in one of the recent Oracle Magazines, and it is available
>online http://www.oramag.com/oracle/98-Jul/dba.html

Are there any ramifications in using the v733_plans_enabled=true parameter this article recommends? Why isn't it enabled by default? Thanks. Received on Thu Oct 22 1998 - 12:27:00 CDT

Original text of this message

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