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 vs Index Full Scan?

Re: Index Fast Full Scan vs Index Full Scan?

From: Vsevolod Afanassiev <vafanassiev_at_aapt.com.au>
Date: 15 Oct 2001 20:57:17 -0700
Message-ID: <4f7d504c.0110151957.4b28534e@posting.google.com>


To do "Index Fast Full Scan", Oracle reads index blocks in no particular order - simply in the order it can find them in the datafiles. Probably (not sure) it reads multiple blocks in a single read, just like during full table scan (using DB_FILE_MULTIBLOCK_READ_COUNT - for example, on Sun platform it can read 1 MB in a single read). For "normal" index reads Oracle starts from the root block, then goes to branches, and then to the leaf blocks - this way it can return results in sorted order.

I don't know what is "Index Full Scan" - where did you find it?

Regards,
Sev

vafanassiev_at_aapt.com.au

sweidanz_at_yahoo.com wrote in message news:<qbJy7.5$mL2.511_at_nsw.nnrp.telstra.net>...
> I dont totally understand whats the difference between these two access
> methods? I know they both access the index only to get the requested
> records because all the columns specified in the query including the
> predicate clause exist in the index. But the thing that i dont understand
> is the differences and why the index fast full scan does not return the
> data sorted?
>
> Thanks,ZS
Received on Mon Oct 15 2001 - 22:57:17 CDT

Original text of this message

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