Re: adaptive cursor sharing and bind peeking
- From: Mohamed Houri <mohamed.houri@xxxxxxxxx>
- To: Jonathan Lewis <jonathan@xxxxxxxxxxxxxxxxxx>
- Date: Thu, 30 May 2013 14:12:53 +0200
According to the tests I have done, I am inclined to say that Carlos
description is correct. If I have already a child cursor that is bind
sensitive, bind aware and shareable for a bind variable that favors (for
example an index range scan) and then I execute the same query with a
different bind variable which is also favoring the same index range scan,
It will share the same child cursor. I have to test that again but I am
quite confident that it will work as described.
However, I have found situations where this smooth ACS working is disturbed
http://hourim.wordpress.com/2013/04/04/what-can-impeach-adaptive-cursor-sharing-kicking-off/
Best regards
Mohamed Houri
2013/5/30 Jonathan Lewis <jonathan@xxxxxxxxxxxxxxxxxx>
>
> Carlos,
>
> This doesn't seem to agree with a blog post that Randolf Geist wrote some
> time ago:
> http://oracle-randolf.blogspot.co.uk/2011/01/adaptive-cursor-sharing.html
>
> He does mention a bug, relating to ACS not working properly with PL/SQL,
> and presumably that's fixed by now; even so, "checking before every
> EXECUTION" doesn't seem to match his observations.
>
> Regards
> Jonathan Lewis
>
>
> ________________________________________
> From: oracle-l-bounce@xxxxxxxxxxxxx [oracle-l-bounce@xxxxxxxxxxxxx] on
> behalf of Carlos Sierra [carlos.sierra.usa@xxxxxxxxx]
> Sent: 30 May 2013 11:35
> To: ax.mount@xxxxxxxxx
> Cc: Christian Antognini; oracle-l@xxxxxxxxxxxxx
> Subject: Re: adaptive cursor sharing and bind peeking
>
> Once it is bind aware on every execution it looks at the values of binds
> ...--
> //www.freelists.org/webpage/oracle-l
>
>
>
--
Bien Respectueusement
Mohamed Houri
--
//www.freelists.org/webpage/oracle-l
Other related posts:
