Re: Special Characters Created on a UNIX system with HP Terminal on US7ASCII

On Jul 18, 10:40 pm, "Martin T." <bilbothebagginsb..._at_freenet.de> wrote:
> On Jul 17, 3:46 pm, gk24 <g..._at_nurfuerspam.de> wrote:
>
> > Hi all,
>
> > I have a customer who uses an application on a UNIX server. The
> > database (US7ASCII, 9i) is on a Windows server. The system was
> > migrated from a 2-tier solution before meaning the application and the
> > database were on UNIX before.
>
> > (...)
> > The nls_lang settings are US7ASCII on all three tiers. The dump
> > function shows the e.g. the German "ü" is stored as &CF in the
> > database.
>
> [Seems Google ate my reply =|]
>
> Something is not right here: You are using a US7ASCII DB to store
> German Umlauts? I wonder how that ever worked.
> Character code for ue" is 0xFC so that looks about right.

It works because some tools and applications don't check NLS. It breaks because some tools do check it. So Frank's suggestion of setting NLS only on Windows is correct for the short term. In the long term, it is a time bomb waiting to go boom, because exp/imp, for example, will oh-so-helpfully translate between character sets if they are set different.

Previous threads here by Frank, and several notes on metalink have clues how to fix this. Not out of luck, just need to work it out. There is a character conversion utility which includes a way to figure out how bad the problem is. But Gerrit, and anyone else using US7ASCII should fix the database character set sooner rather than later.

>
> Maybe, in addition to looking up the stuff in the Oracle Globalization
> Manual you also want to look up ISO-8859-1 (Latin1) and Windows-1252
> (CP1252)
>
> br,
> Martin

jg

--
@home.com is bogus.
http://cleanishappy.com/