Try it on SO level (unix): you can order file (for instance, on the first
field and delimiter is comma '$sort -t, -1 example.dat > example.sort') and
then extract unique register (for instance, '$uniq example.sort >
example.uniq')... the result could be a good file to load into DB.
Ciao


-----Mensaje original-----
De: Sikyala Raquel [USA]
[mailto:oracledba-ezmlmshield-x73241659.[Email address protected]
Enviado el: viernes, 28 de marzo de 2008 21:54
Para: LazyDBA Discussion
Asunto: Is there a way within sqlldr to delete any duplicates found during
the load

I have created a cron job to run sqlldr in a ksh script against text
flat files on a daily basis. It runs in a loop because on any given day
there may be more than one file to process. Initially I was thinking
that the table it loads into should be truncated prior to the data load
because just in case there were duplicate records loaded. However, after
thinking about it I think that there is a potential to lose data. Is
there a way within sqlldr to delete any duplicates found during the load
in the table? The control file I am using currently does an append into
the table.



No virus found in this outgoing message.
Checked by AVG.
Version: 7.5.519 / Virus Database: 269.22.1/1350 - Release Date: 30/03/2008
12:32



Oracle LazyDBA home page