Re: v$access oddity?
From: Mladen Gogala <gogala.mladen_at_gmail.com>
Date: Fri, 18 May 2012 12:22:22 +0000 (UTC)
Message-ID: <pan.2012.05.18.12.22.22_at_gmail.com>
On Thu, 17 May 2012 16:05:12 -0700, Noons wrote:
> TYPE -----------
> 95 DWDEVTEAM F_GL_MONTHLY_BALANCE
> TABLE
> 95 DWDEVTEAM F_GL_MONTHLY_BALANCE
> MULTI-VERSI ONED OBJECT
>
>
>
> F_GL_MONTHLY_BALANCE is a partitioned table.
> "Multi-versioned object"?????
> Looked it up everywhere, doesn't seem to register in google.
>
> Anyone seen this before? What exactly is a "multi-versioned object"?
> In this context, of course: I'm fully aware of UNDO mechanism and
> Oracle's multi-versioning change control.
> No other users of this table other than my SID above (95).
> And I am not changing it, I'm re-calcing stats.
Date: Fri, 18 May 2012 12:22:22 +0000 (UTC)
Message-ID: <pan.2012.05.18.12.22.22_at_gmail.com>
On Thu, 17 May 2012 16:05:12 -0700, Noons wrote:
> This, from a query on v$access in our DW in 11.2.0.3.2:
>
> SID OWNER OBJECT
> -------------- --------------------
>
> TYPE -----------
> 95 DWDEVTEAM F_GL_MONTHLY_BALANCE
> TABLE
> 95 DWDEVTEAM F_GL_MONTHLY_BALANCE
> MULTI-VERSI ONED OBJECT
>
>
>
> F_GL_MONTHLY_BALANCE is a partitioned table.
> "Multi-versioned object"?????
> Looked it up everywhere, doesn't seem to register in google.
>
> Anyone seen this before? What exactly is a "multi-versioned object"?
> In this context, of course: I'm fully aware of UNDO mechanism and
> Oracle's multi-versioning change control.
> No other users of this table other than my SID above (95).
> And I am not changing it, I'm re-calcing stats.
Nuno, statistics is the multi-versioned object, not the table itself. I
believe that you will see something completely different if you did
"select count(*) from F_GL_MONTHLY_BALANCE" and I don't have in mind John
Cleese in a bikini.
Stats on version 11G can have several versions, one being used, another
one being QA-ed, the third one being collected.
-- http://mgogala.byethost5.comReceived on Fri May 18 2012 - 07:22:22 CDT