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: index fast full scan - why so slow and expensive

Re: index fast full scan - why so slow and expensive

From: Anurag Varma <avdbi_at_hotmail.com>
Date: Wed, 02 Feb 2005 06:05:48 GMT
Message-ID: <05_Ld.17495$8H2.16018@twister.nyroc.rr.com>

"Sybrand Bakker" <sybrandb_at_hccnet.nl> wrote in message news:h9p001dloosmvs7i5qdacl2v6i327lfesc_at_4ax.com... -snip-
> - The INDEX hint will use the normal access methods. However it is a
> *hint*. As you are deliberately vague about the actual statement, it's
> explain plan etc, and you obviously didn't run explain plan with event
> 10053 set, it is impossible to tell why Oracle choose the index_ffs
> method
>
> - I hardly use it, you will need to very that in the docs
>
> - If the clustering factor approximates the number of records it is
> good (recalling from memory), it is approximates the number of blocks
> it is bad.

opposite. If CF nears number of table blocks .. its good

               If CF nears the number of table records .. its bad.

> You would need to rebuild the table and make sure the records are
> loaded in the correct order. Obviously, if your table is a heap table,
> you would need to repeat this periodically.
>
>
> --
> Sybrand Bakker, Senior Oracle DBA

To OP ... have a look at www.dizwell.com -> DBA Tips -> Tuning -> The clustering factor ... for an excellent writeup of clustering factor by Howard.

Anurag Received on Wed Feb 02 2005 - 00:05:48 CST

Original text of this message

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