Hi,

I have written a C application using OCI8.1.7 API to work with Oracle
Unicode Data Types NVARCHAR2. On executing the application, I receive
following error

ORA-03113: end-of-file on communication channel.

But the same application works with OCI9. Does that mean OCI8 does not
support 'N' data types? If it does then what is reason of the above
error.

Any response in this regard will be helpful.

On Apr 17, 1:20 pm, sachin_...@yahoo.com wrote:
> Hi,
>
> I have written a C application using OCI8.1.7 API to work with Oracle
> Unicode Data Types NVARCHAR2. On executing the application, I receive
> following error
>
> ORA-03113: end-of-file on communication channel.
>
> But the same application works with OCI9. Does that mean OCI8 does not
> support 'N' data types? If it does then what is reason of the above
> error.
>
> Any response in this regard will be helpful.

Oracle has altered national characterset names, and reduced the number
of national charactersets in 9i. I would recommend looking in the
globalization manual for both versions, whether your characterset is
valid in both versions.
v$nls_valid_values will show you which values are possible.
Apart from that ora-3113 can point to almost anything and requires a
search on Metalink.
Note that 8.1.7 is desupported for quite some time, and 9i will follow
soon.

--
Sybrand Bakker
Senior Oracle DBA

On Apr 17, 4:46 pm, sybrandb <sybra...@gmail.com> wrote:
> On Apr 17, 1:20 pm, sachin_...@yahoo.com wrote:
>
> > Hi,
>
> > I have written a C application using OCI8.1.7 API to work with Oracle
> > Unicode Data Types NVARCHAR2. On executing the application, I receive
> > following error
>
> > ORA-03113: end-of-file on communication channel.
>
> > But the same application works with OCI9. Does that mean OCI8 does not
> > support 'N' data types? If it does then what is reason of the above
> > error.
>
> > Any response in this regard will be helpful.
>
> Oracle has altered national characterset names, and reduced the number
> of national charactersets in 9i. I would recommend looking in the
> globalization manual for both versions, whether your characterset is
> valid in both versions.
> v$nls_valid_values will show you which values are possible.
> Apart from that ora-3113 can point to almost anything and requires a
> search on Metalink.
> Note that 8.1.7 is desupported for quite some time, and 9i will follow
> soon.
>
> --
> Sybrand Bakker
> Senior Oracle DBA

Probably my query was incomplete.... My previous statment remains
same, I just want to add that I am connecting to Oracle10g server
using my C application written by using OCI8.1.7 client. In Oracle10g
server by default AL16UTF16 encoding is used for NVARCHAR2 data type.
I have checked the Oracle documentation that AL16UTF16 encoding is
supported from Oracle9.0.1 onwards, So Oracle8.1.7Database does not
support AL16UTF16. But does that mean Oracle8.1.7 client can not be
used to insert Unicode data in 'N' data type of Oracle 10g Server?

On 17 Apr 2007 05:33:50 -0700, sachin_mzn@yahoo.com wrote:

>On Apr 17, 4:46 pm, sybrandb <sybra...@gmail.com> wrote:
>> On Apr 17, 1:20 pm, sachin_...@yahoo.com wrote:
>>
>> > Hi,
>>
>> > I have written a C application using OCI8.1.7 API to work with Oracle
>> > Unicode Data Types NVARCHAR2. On executing the application, I receive
>> > following error
>>
>> > ORA-03113: end-of-file on communication channel.
>>
>> > But the same application works with OCI9. Does that mean OCI8 does not
>> > support 'N' data types? If it does then what is reason of the above
>> > error.
>>
>> > Any response in this regard will be helpful.
>>
>> Oracle has altered national characterset names, and reduced the number
>> of national charactersets in 9i. I would recommend looking in the
>> globalization manual for both versions, whether your characterset is
>> valid in both versions.
>> v$nls_valid_values will show you which values are possible.
>> Apart from that ora-3113 can point to almost anything and requires a
>> search on Metalink.
>> Note that 8.1.7 is desupported for quite some time, and 9i will follow
>> soon.
>>
>> --
>> Sybrand Bakker
>> Senior Oracle DBA
>
>Probably my query was incomplete.... My previous statment remains
>same, I just want to add that I am connecting to Oracle10g server
>using my C application written by using OCI8.1.7 client. In Oracle10g
>server by default AL16UTF16 encoding is used for NVARCHAR2 data type.
>I have checked the Oracle documentation that AL16UTF16 encoding is
>supported from Oracle9.0.1 onwards, So Oracle8.1.7Database does not
>support AL16UTF16. But does that mean Oracle8.1.7 client can not be
>used to insert Unicode data in 'N' data type of Oracle 10g Server?
>
>

This question can't be answered, as you always need to provide all 4
digits of a version.
Only 8.1.7.*4* can connect to 10g (and is equally desupported as other
versions of 8i)
As far as I know the information in v$nls_valid_values is derived from
the client. If you have a 8.1.7 sqlplus. you should be able to see
which charactersets are supported.