Re: Converting from WE8ISO8859P1 to UTF8

I'm surprised the import wouldn't run at all without data. If you are importing to a UNIX box you need to set an environment variable NLS_LANG even if all your NLS parameters are set perfectly in when you install Oracle and in your init.ora file when you create the database.

Van

<ovema_at_asp.infostream.no> wrote in message news:rpkem326d2h.fsf_at_vippetangen.asp.infostream.no...
>
> Hi.
>
> I need to convert a already built schema from WE8ISO8859P1 to UTF8.
>
> 8859-1 is not a strict subset of UTF8 so converting directly is not
> possible. All the characters in 8859-1 are in the UTF8 set so there
> should be possible to convert without dataloss if done correctly.
>
> My current plan is to recreate the tables, indexes etc on the new
> database then make a perl-script to move the data. I have never tried
> using DBI with anything else than 8859-1 but that shouldnt be to big a
> problem.
>
> So over to the problem:
>
> I dont have the build-scripts for the schema.
>
> Is there anything out there that can read the schema from one database
> and generate the needed SQL to make the same schema on a new database?
> I would hate to have to go trough the database manually to get all the
> information.
>
> I tried running exp/imp with rows=n but import still refuses to run
> due to the different char-sets.
>
> --
> Ove
Received on Sun Sep 03 2000 - 08:48:30 CDT