To me what is unsettling is having an enabled constraint with an unusable index; one would expect if the index is unusable the constraint would be invalid. David Fitzjarrell From: "Bobak, Mark" <Mark.Bobak@xxxxxxxxxxxx> To: "info@xxxxxxxxxxxxxxxxxxxxxxxxx" <info@xxxxxxxxxxxxxxxxxxxxxxxxx>; "oracle-l@xxxxxxxxxxxxx" <oracle-l@xxxxxxxxxxxxx> Sent: Wednesday, November 2, 2011 9:49 AM Subject: RE: Can t get INSERT /*+ APPEND */ to indicate direct load in execution plan Understood....it's just a bit unsettling that you can have an *enabled* constraint, and still get duplicate data into the table. -Mark -----Original Message----- From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx] On Behalf Of Randolf Geist Sent: Wednesday, November 02, 2011 12:42 PM To: oracle-l@xxxxxxxxxxxxx Subject: Re: Can t get INSERT /*+ APPEND */ to indicate direct load in execution plan Hi Mark, good question - it's quite a while ago - and I think to remember since this was not part of a client's problem I didn't open a SR. Note though that the duplicates are only possible when the non-unique index supporting the PK is in unusable state. Under normal circumstances everything is fine. Randolf > Wow, that's a nasty one, Randolf. > > Did you open SR with Oracle on this? > > If so, what did they say? > > > Nasty..... > > > -Mark -- //www.freelists.org/webpage/oracle-l -- //www.freelists.org/webpage/oracle-l -- //www.freelists.org/webpage/oracle-l