Re: sqlldr doesn't convert unicode to WE8ISO8859P1

Owain,

I think Oracle can only perform conversions between character sets where the source set is a subset of the destination set. So ASCII could be converted to Unicode but not vice versa.

If your data is just plain ASCII characters encoded as unicode, could you use an operating system tool to perform the conversion? For example, would loading the data into an application like notepad then saving it as plain ASCII work?

Alternatively, could you define the NLS character set for your database as UTF8 then load the data into nchar columns?

Regards,

Richard

"Owain Cole" <owain.cole_at_xenicom.com> wrote in message news:7a3109e1.0204160446.6f834cf5_at_posting.google.com...
> Hi,
>
> My database is defined to use characterset WE8ISO8859P1 and I need to
> load some data which has come from an Excel spreadsheet, converted to
> unicode tab delimited format.
>
> The below sqlldr control file loads the unicode into the required
> table, however it doesn't appear to have been converted to
> WE8ISO8859P1 as the data looks a little like this ' h e l l o w o r l
> d'.
>
> here's the script (I'm a newbie to sqlldr so bare with me ;)
>
> LOAD DATA
> CHARACTERSET UTF8
> INFILE unicode.txt
>
> BADFILE unicodebad.txt
>
> REPLACE
>
> INTO TABLE PERMITS
> FIELDS TERMINATED BY X'09' OPTIONALLY ENCLOSED BY '"'
> TRAILING NULLCOLS
> (
> Name CHAR,
> SITE_OPTION CHAR
> )
>
> And the command line is:
> D:>sqlldr user/pswd_at_schema unicode_loader.ctl -log log
> .txt
>
>
> Please e-mail me any thoughts on this (our it policy is time
> restrictive wrt internet access)!
>
> Thanks,
> Owain
Received on Tue Apr 16 2002 - 08:18:09 CDT