Re: High Version count in V$SQLAREA

srivenu_at_hotmail.com (srivenu) wrote in message news:<1a68177.0307092113.619f62a0_at_posting.google.com>...
> I really appreciate everyone who has answered and especially yong.
> I think, yong, you are right.
> The high version count is due to Bind Mismatch.
> I gave the following query and got the answer for my question.
>
> sys_at_OMPROD> l
> 1 select substr

...
> 25 from v$sql a,v$sql_shared_cursor b
> 26 where a.address=b.kglhdpar
> 27 and sql_text like '%DATE_ENTERED,NOTES_ID%'
> 28* order by 2
> sys_at_OMPROD> /
>
> Par Loa
> First sing Reason For ded
> Open Kept Users Users
> Child Load User Cursor Not Vers
> Vers Vers Ope Execu
> SUBSTR(SQL_TEXT,1,25) Number Time id Being Shared ions
> ions ions ning ting
> ------------------------- ------ ---------- ----- --------------- ----
> ---- ---- ----- -----
> INSERT INTO ASAP.NOTES ( 0 2003-06-21 0 -Bind Metadata 0
> 0 0 0 0
> /02:00:34 Mismatch-
>
> INSERT INTO ASAP.NOTES ( 1 2003-06-21 0 -Bind Metadata 0
> 0 0 0 0
> /02:00:34 Mismatch-
>

...
>
>
> 32 rows selected.
>
> Now these statements are coming from an application.
> Can someone tell me how i can rectify this ?
> Thanks in advance

Hi, Srivenu,

Anurag suggests setting _sqlexec_progression_cost to 0. Can you try that?

BTW, you didn't explain why you want to thank me more than others in this thread. All I did out of the thread is ask James Morle, and he replied that my interpretation of his book is correct and he strongly suspects bind threshold alone may explain why you have so many versions (before he sees your result from v$sql_shared_cursor).

Yong Huang Received on Thu Jul 10 2003 - 12:17:46 CDT