Guys, Thanks. Say a table spans four datafiles, rowid being the physical address of the row, comparing the rowid of one row in datafile 1 with the rowid of another row in datafile 4 did not make sense to me. I think it is kind of comparing my street address in CA with Jared's street address in OR. :-) Cheers! Arul On 11/29/06, Niall Litchfield <niall.litchfield@xxxxxxxxx> wrote:
I agree with Jared that different results looks like a bug, I also agree however with Arul that the comparison operation seems meaningless. It would seem in the same class as < comparisons for colours On 11/28/06, Jared Still <jkstill@xxxxxxxxx> wrote: > On 11/28/06, Arul Ramachandran <contactarul@xxxxxxxxx> wrote: > > > > Oracle 10.2.0.1: > > > > > To me comparing two rowids does not make sense, I can't see how one rowid > > be greater than another? But then I see a ML note 2714430.8: "Bug 2714430 > > - Wrong results from ROWID comparison to rowid with slot#>32767" that > talks > > about rowid comparison. > > > > Any insights on this? > > > > > > I thought perhaps the data type for a ROWID in a table might > differ from the data type used for a transient ROWID. > (this happens with DATE types) > > Not so. > > 14:32:23 SQL>l > 1* select dump(rid1), dump(rowid) from dummy > 14:32:24 SQL>/ > > DUMP(RID1) > ---------------------------------------------------------------------------------------------------- > DUMP(ROWID) > ---------------------------------------------------------------------------------------------------- > Typ=69 Len=10: 0,0,163,228,2,129,27,59,0,2 > Typ=69 Len=10: 0,0,215,37,1,0,1,14,0,0 > > > 1 row selected. > > Both the same type. Sure looks like a bug. > > -- > Jared Still > Certifiable Oracle DBA and Part Time Perl Evangelist > > -- Niall Litchfield Oracle DBA http://www.orawin.info