We use proC. we are in DB version 10.2.0.5. I do not have date stamps on the tables that have the issue, but by joining, to other tables, it looks like at least 1 record was created in the last 6-12 months. When I run dump(col,1016), I get Typ=2 Len=3: c0,0,0 note that my rawtohex value is C0000, so it appears related. Anyone know waht c0,0,0 means? Or where I can look it up. I am guessing Ill find its an invalid encoding. I take it this means garbage data. Is there any easier way to look for bad data without running a dump on all fields in the database? We are not using RMAN. We do straight EMC SAN level BCV backups. I don't think the RMAN utility would help. I don't think this is block corruption since I can actually query the data. On Fri, May 4, 2012 at 4:24 PM, Rich Jesse <rjoralist2@xxxxxxxxxxxxxxxxxxxxx > wrote: > Jeffrey Beckstrom writes: > > > I vaguely recall that a very long time ago there was an issue where some > > PRO*x programs could insert bad data into the database without being > caught. > > That's all I recall about it at the moment. > > The only one of those I know of -- nay, lived through -- was that a numeric > zero could be inserted into a date datatype column, which caused all sorts > of fun. > > Even though it was introduced in v6 or v7, I've never seen documentation > that the bug has been closed, supposedly due to its widespread (ab)use. > I've experienced the bug in v9 and possibly v10. > > Rich > > -- > //www.freelists.org/webpage/oracle-l > > > -- //www.freelists.org/webpage/oracle-l