Sybrand Bakker <postbus_at_sybrandb.demon.nl> wrote in message news:<22787us33nt2e4sptcc28pgtp5vlk2iq3m_at_4ax.com>...
> On Wed, 20 Feb 2002 18:22:03 +0100, Svend Jensen
> <Master_at_OracleCare.Com> wrote:
>
> >Tianhua Wu wrote:
> >
> >> , some consistent lookups in a
> >>
> >>>rollback segment was just deallocated by optimal, just before you got
> >>>there - the database engine has no problems finding the data, they are
> >>>there, well preserved in a deallocated rollback segment.
> >>>
> >>
> >> Are you sure about that? That is contradict to my experience. I
> >> understand that the info is still there. However Oracle will use a
> >> block that does not belong to the segment anymore? I definitely need
> >> proof for this claim. If this is true, you should have no trouble get
> >> data back right after you truncate a table as long as nobody else use
> >> the space (Oracle can just use similar mechanism)! Of course you can
> >> say flash back query for 9i, but that is different story.
> >> Unfortunately I do not have a testing database now.
> >>
> >
> >Truncate doesn't generate rollback, and yes the data is still there on
> >the file on disk, but you can't get to it with std. oracle ways.
> >And i'm not sure all versions on all platform behave alike, so a test is
> >needed with your version on your platform.
> >/svend
>
>
> Truncate is DDL, so it is committed automatically
>
> Regards
>
> Sybrand Bakker, Senior Oracle DBA
>
> To reply remove -verwijderdit from my e-mail address
Yes, I do not question truncate do committing at all. I would assume
so it is when you shrink the rollback segment (it need to update dict
tables and commit). Otherwise everybody would still think the blocks
still belong to the original RBS.
Received on Thu Feb 21 2002 - 10:45:03 CST
