Strange unicode problem with oracle 10g
From: Sascha Schroeder (sascha.schroeder_at_e-pro.de)
Date: 03/17/05
- Previous message: Del M: "Re: Building master detail relationship"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Thu, 17 Mar 2005 14:41:55 +0100
Hello,
we have a strange problem with unicode characters and oracle 10g. If i
insert a unicode character (0x88AF)
into a oracle varchar field i get the following error:
ORA-01460: unimplemented or unreasonable conversion request.
Ok, let's go on to the very strange things. Now i insert the same unicode
character and an additional
single byte character in one literal, for example 'a' (0x61). Now everything
works fine. The data is stored
in my table.
I've figured out that i get an error if my literal contains more multibyte
(mb) characters than
single byte (sb) characters. if number of sb chars >= number of mb chars it
works.
That's the behavior if i use oracle 10g server and oracle 10 ole db
provider. But if i use
oracle 10g server with oracle 8 ole db provider on my client everything is
fine.
I've tested the complete scenario with delphi ado components and imported
ado type library.
The error occurs only with delphi components, with ado type library i have
no problems.
We have also tested with perl and ado type library and it works too. I think
the error is somewhere
in the delphi wrapper for ado.
At the moment we are installing Oracle 8 client at our customers as a
workaround, but that's not
a final solution for the problem.
Any ideas what kind of problem i have?
Thanks a lot,
Sascha
-----------------------------------------------
Environment:
Client:
- Windows 2000
- Delphi 7
- Oracle Client (8 or 10)
Server:
- Windows 2000
- Oracle 10g
- Database in UTF-8
- Target field is varchar. Length 255
- Previous message: Del M: "Re: Building master detail relationship"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|