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 vs. table scans in statspack reports

Re: Index vs. table scans in statspack reports

From: Tanel Poder <change_to_my_first_name_at_integrid.info>
Date: Sun, 30 Nov 2003 18:45:15 +0200
Message-ID: <3fca1e9c$1_1@news.estpak.ee>


> I just read an article at:
>
> http://www.optimaldba.com/internals/oraint_dual.html
>
> There, a new structure, a view VDUAL to x$dual is proposed to aliviate
> the access cost to DUAL. The author does not dare to substitute the
> original DUAL view, though. It's a pity not to know if it would work
> without having to reedit all queries using DUAL.

You can create a local dual synonym under your user's schema, which points to v_dual view (on x$dual table) under sys schema. Local synonyms are always preferred over public ones, so allowing you to use "fast" dual without modifying your code or recreating original dual.

And if you forget to create local dual synonym under some users, Oracle just uses the original dual through public synonym. Personally, I would add an

Tanel. Received on Sun Nov 30 2003 - 10:45:15 CST

Original text of this message

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