Re: sqlldr and discarded records

Lothar Armbruester wrote:
> Christoph Seidel wrote:
>> Alistair Thomson wrote:
>>> Hi
>>>
>>> Use the .bad file as the new source file, after you've resolved the
>>> issues why the records weren't posted the first time around.
>>
>> but what about the imput file? some records were already imported,
>> some not
>> (not in bad file).
>> How do i prevent that the already imported files get re-imported?
>>
> Hello Christoph,
>
> in your loader control file you can specify a badfile and a
> discardfile the latter receiving the discarded rows.
> So after your load you can process the discardfile instead of your
> input file. Obviously you have to adust your loader control file or
> the rows will be discarded again.
> Since the discardfile only contains discarded rows, there is no danger
> to load some rows twice.

well it seems, i do not understand, what discarded means.. :/

imo:

bad: all what is rejected due to oracle errors (bad format, length, etc) discarded: all what does not match a loading filter (when)

but what i meant was the bad file.

but nevertheless, your advice does not handle the case where the input file was not process completely (maybe the loader dies for some reason):

rejected records are in the bad file, fine, i can fix it and do the import with this file as input file

not processed records AND imported records are still in input file, do i have to split now the imput file into imported and non-imported? Received on Thu Mar 13 2003 - 11:28:16 CST