Welsh Fonts in Oracle database

[Big Mac]

data pasted into an oracle database from word or outlook contains welsh fonts
with caret after translating a paragraph into welsh. This automatically
changes y or w with caret (unicode 0176 abnd 00FB resp) into other symbols.
they have no alt code alternatives. y and w end up as a | and appostrophes
turn in to upside down ?.

The font in the source is arial but the database appears to be something
like Berlin Sans FB. it is a 3rd party database and we can't open it in
design mode.

Does anyone know what's happening and can you tell me what we need to do to
resolve this as it is a critical departmental database.

Thanks in advance for your time on this matter

 

[Gordon Bentley-Mix on news.microsoft.com]

This appears to be a problem on the database side of the process. As such,
it's doubtful anything you could do in Word would help. I would recommend
contacting the database developer for a solution.
--
Cheers!

Gordon Bentley-Mix
Word MVP

Please post all follow-ups to the newsgroup.

Read the original version of this post in the Office Discussion Groups - no
membership required!

 

[Peter Jamieson]

I would agree with Gordon that this is likely an Oracle issue, and I
don't currently have a copy here, but a few observations...
a. despite the fact that you cannot see the design, is it possible to
determine what character set the Database was created to use? (i. how
this works may have changed in really recent versions of Oracle; ii.
If it isn't a Unicode set, or it is not based on ISO 8859-14, that is
likely the problem. Even if it is based on ISO 8859-14, you will likely
have problems with any other encodings you need.
b. Then there is the issue of the character set of the "client", which
in this case means all the relevant stuff on the PC, but more
particularly, whatever software components are transporting data to/from
your Oracle front-end. AIUI the client software can be set up
differently and a non-Unicode character set, or a different encoding
from the server, will cause you problems. I would guess, but do not know
for sure, that any locale settings on your PC do not actually affect
that setting.
c. then there's the question of what Word is putting into the Windows
clipboard when you copy, and what your oracle front end is taking from
the clipboard when you paste. Typically, Windows applications either
paste multiple different formats into the clipboard, and/or declare
other formats they are willing to render on demand. So for example Word
might provide plain text, RTF etc.: the Orcale client will have a
mechanism for choosing which format to paste (e.g. it probably wont use
RTF). It may even be worth looking for a "paste special" option in the
Oracle client to see if you can alter the way it pastes. But if Word
provides a non-Unicode text encoding and Oracle takes that one, there's
your problem.
d. and finally there's the question of character set support in the
font used by your Oracle client (as you have already suggested).

I suppose one thing you can easily confirm/deny is whether it's just
Word that's causing the problem - if you start with those characters in
Notepad, for example, do you have the same problem?

This automatically
changes y or w with caret (unicode 0176 abnd 00FB resp)

BTW AFAICS the uppercase versions were 0176 and 0174 and 00FB was a u
with caret, so if youi really have w with caret at 00FB I'd say
something else needs looking at as well.


Peter Jamieson

http://tips.pjmsn.me.uk