Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.misc -> Re: how Oracle processes index scans

Re: how Oracle processes index scans

From: Dogan Cibiceli <dcib_at_gwl.com>
Date: Fri, 20 Aug 1999 19:53:50 -0600
Message-ID: <37BE06AE.CF7A715F@gwl.com>


Steve Perry wrote:

> Does anyone know how Oracle processes a range scan on an index? Does it
> find the high and low values in the index and work toward the center
> or???
> What about other types of index access?
>
> Thanks,
> steve

As far as I know Oracle uses B++ tree indexing mechanism in general if you donot specify. Well in a range scan, the problem is that the address on the file will not be continous. So oracle will probably recursively descend down the B++ index and get all the file address (rowids). From thereon, I donot know whether it does an internal sort of rowid so that it doesnot have to access a single block multiple times ? Therefore, a range scan isnot a fast method of searching data unless your data is sorted on the disk and you can do a single sweep with one disk access and read the stuff.

Dogan

PS. This is highly guesswork... Received on Fri Aug 20 1999 - 20:53:50 CDT

Original text of this message

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