Re: extended ASCII seems not to exist for some clients

Christian BIERO wrote:
> Hello,
>
> I've got the following problem.
> I've got a script to insert data into some table. This script contains some
> accents.
> The target database has the following DATABASE params (cf select below)
> According to the charset : WE8ISO8859P15, the DB supports accents like 'é'
> or 'à'.
>

<snip!>

This has little to do with the database, but with the charactersets defined on server and client side. Actually, the network layer is responsable for the "translation" between different charactersets. If your client is US7ASCII (which used to be the default Oracle characterset for far too long!), no extended characters whatsoever! Could it be that:
- you server/telnet session does not have a characterset set at all?   (check your environment)
- you DOS session uses codepage 850 (as said before).

Your é has code 0xE9; that is é translated to 0xE9 in cp 8859P15 (Iso Latin 1 plus Euro). In cp850 (DOS), 0xE9 is Ú. Defined nothing, the client would probably use US7ASCII. As 0xE9 is not supported (it's 8 bits), the msb falls off, resulting 0xE9 to become 0x69. This is the letter i.

Works for me...

BTW, 8859P15 will not process the euro sign correctly in Windoze environments, M$ has chosen, in all it's wisdom, to map the Euro to an entirely different character than the ISO - use MSWIN1252 instead.

Grtz, Frank Received on Thu Dec 19 2002 - 03:06:33 CST