Re: value from index block or table block

From: Subodh Deshpande <deshpande.subodh_at_gmail.com>
Date: Wed, 23 Nov 2011 23:01:48 +0530
Message-ID: <CAJsOtB6N=37O8O9H6AO0gaFzRpFozLJw71ng_pbuQfo36A+XSg_at_mail.gmail.com>

Hello,
in your case index is not getting used..i suppose...

you mean ..you created a table and some indexes..block corruption is on index and you fired query..
it should error out in this case...

index is not meant to give the subset (fulfilled by query) of data BUT to present it in the manner you created the index...the query clause is going to give the subset of data...

thanks...subodh

On 23 November 2011 21:26, Grzegorz Goryszewski <grzegorzof_at_interia.pl>wrote:

> Hi,
> I know that seems crazy but please clarify that for me ,let say for
> 10.2.0.3 but that does not matter .
>
> Let say we got table t with col1, col2, col3 and index on that table on
> col1, col2 and
> query
>
> select col1,col2,col3 from t where col1 = 'value' ;
> Let say that
> TABLE ACCESS BY INDEX ROWID was used for that query
>
> My question from where Oracle retrieves col1,col2 values ?
> From index block and then col3 value from table block via rowid found in
> index block ?
> Or maybe only rowid from index and then col1,col2,col3 values from table
> block only ?
>
> I'm asking because wonder if its possible to get different values if
> access plan is full table scan .
> Suppose You've got 'corrupted' index block (or rather changed by block
> editor) with different value than in table block for some column value .
> So it is possible to get different values if access path differs ?
> Regards
> GregG
>
> --
> http://www.freelists.org/webpage/oracle-l
>
>
>

-- 
=============================================
TRUTH WINS AT LAST, DO NOT FORGET TO SMILE TODAY
=============================================


--
http://www.freelists.org/webpage/oracle-l

Received on Wed Nov 23 2011 - 11:31:48 CST