Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> NLS Question - Oracle 8.1.5, UTF8 and Nchar, Nvarchar2, Nclob

NLS Question - Oracle 8.1.5, UTF8 and Nchar, Nvarchar2, Nclob

From: Mark Tompkins <tompkins_at_digital-dispatch.com>
Date: Thu, 27 May 1999 09:03:47 -0700
Message-ID: <374D6CE2.EAD1ADD1@digital-dispatch.com>


Hi,

Oracle 8.1.5 on NT SP 3. 256 MBytes RAM. OFA database. Two instances

Oracle 8.1.5 documentation indicates (correct me if I am wrong) that multi-byte character set support is only possible for Database National Character sets (as opposed to the Database Character set, which only allows single byte character sets). As far as I can see, Oracle only supports UTF-8 variable length character set.

  1. I was able to build a database with UTF-8 set for both the database character set and database national character set. I don't know what the parameter for database character set is (please tell me what it is),

and I can't see the database national character set parameter (NLS_NCHAR) using OEM (why is it hidden, and where is it hiding?). What

I would expect to see is that the N* char types would support multi-byte

characters for UTF-8. Can't do an insert into a table that has a column

that is nvarchar2(255). Why does Oracle let me create the table, if it won't allow me to use the field in the table? Is there some weird configuration setup that needs to be implemented before I pass go to collect my $200?

2. I can't convert from a varchar2 -> nvarchar2 (at least this is documented). However, why can't I? Can't assign a value to the nvarchar2 field any way I try. Does this mean that the Oracle 8.1.5 documentation has jumped the gun on the product?

3. Java (which uses ucs-2 internall) streams are designed in two ways -

byte based and character based. The byte based streams can handle strings, but convert them to UTF-8 when doing I/O. Java has reader and writer classes to handle character based streams. Readers and writers perform implicit character set conversions (ascii to ucs-2, etc.). How much of this is going to rub off on Oracle 8.x, and if so, when? Right now, only two character sets seem to be supported, and they are only single byte ...

4. What is the performance hit using UTF8 rather than a single byte char set?

What I really need to know is - what can I deliver for unicode support today, in terms of configuration for our development environment, so that our product development team doesn't have to do big alterations 3 months from now when the real unicode support arrives on oracle???

Thanks for your insights in advance!

sigh ...

Mark Tompkins Received on Thu May 27 1999 - 11:03:47 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US