Re: ROWID Query

It would only change if you delete the row and reinsert it. Also you can't be sure that the rowid's are sequential because Oracle could put the row between data you want to keep. Why not give the rows a constant id and just delete where the id=X?
Jim

"Steve Long" <steven.long_at_erols.com> wrote in message news:8rm4i9$4ks$1_at_bob.news.rcn.net...
> rowid is the physical address of a row in a given data block. it can
> change, depending on the operations performed on the data.
>
> you are better off just adding a flag column.
>
> <mark.ragan_at_dial.pipex.com> wrote in message
> news:39de51c1.1978875_at_news.dial.pipex.com...
> > Can anyone clarify the allocation & usage of ROWID's;
> >
> > I have a program which runs frequently and creates in excess of
> > 100,000 records, which I may wish to delete later. There is no unique
> > way of identifying these records, and the table is extremely large so
> > creating an identifying column would take up a lot of wasted space. I
> > know that there are lots of ways of achieving whay I want but would
> > the following idea work???
> >
> > If I know the ROWID of the FIRST record created, and the ROWID of the
> > last ROWID created, am I safe to use the following delete command, or
> > do I risk ROWID's being reused if some of these records have already
> > been deleted?
> >
> > delete from XXXX where ROWID >= FIRST and ROWID <= LAST;
> >
> > TIA
> >
> > Mark
> >
>
>
Received on Sat Oct 07 2000 - 09:55:18 CDT