The upgrade works, thank you all !Gastón Quiroga Allytech S.A. Adrian Klaver wrote: ----- "Gastón Quiroga" <gastonq@xxxxxxxxxxxx> wrote:Thank You Tom: I'll Try to make an update, but the 2 fields are equals row by row, how could I make a difference in the "WHERE" statement? Regards Gastón Quiroga Allytech S.A. Tom Lane wrote: =?ISO-8859-1?Q?Gast=F3n?= <tango@xxxxxxxxxxxx> writes: It's Postgres version 8.0.8 Well, that's pretty ancient, and I see at least one bug in the release history that could result in duplicated rows. I'd counsel an update to 8.0.something-recent. You can probably delete the extra row using a WHERE on ctid. regards, tom lanePer Toms previous post use the following query: select ctid,xmin,xmax,* from pg_shadow; Then use the ctid value of the duplicate value in the where clause. Adrian Klaver aklaver@xxxxxxxxxxx |