Strange behavior with Unicode Program

[rguillen]

Hi support,

I did a software using DOA 4.1.2.0 in Delphi XE, it copy data from Source to Target database.

This software worked perfect when the copy was from Oracle 9i to Oracle 11g R2.

Few days ago the source database was upgraded to Oracle 11g R2, but now there is a strange behavior with some fields. In the version 9 the table structure for some fields were CHAR(2), now the table structure are N CHAR(2), but the program only copy the first character, but the others fields that are nchars(20, or 10) too, the data is copied successful.

This was a migration from JDE One world data to JDE9.
The environment variables for both databases are:
NLS_LANGUAGE: AMARICAN
NLS_TERRITORY: AMERICA
NLS_CHARACTERSET: WE8MSWIN1252.

I did a fast program with Delphi 2007 to check, and it copy successful the data between two database.

Could you please give me an Idea about this issue?

Thanks.

Richard.

 

[Marco Kalter]

This can indeed occur with nchar columns. We'll have a fix available shortly.

 

[rguillen]

I'll waiting this fix as soon as you can will provide me.

Thanks for your quick answer.
Richard.

 

[rguillen]

Any news about this?

 

[Marco Kalter]

You can download a Delphi XE fix here. Let me know if this fixes it for you.

 

[rguillen]

Hi Marco,

I did some test and it appear fixed, but I saw that you give me a Evaluation copy. My service contract expiration date is October 1, 2011.
Do you have the fix for Delphi 2010? Because I have some programs in this version of Delphi.

Thanks.

 

[Marco Kalter]

Okay, we'll upload a final version shortly.

 

[rguillen]

Hi Marco, any news about this fix. I really need it.

 

[Marco Kalter]

The finale registered Delphi XE version can be downloaded here. A Delphi 2010 (and all other versions) will follow shortly.

 

[rguillen]

Hi Marco,

Thanks by your help. I'm a little bit confuse. I have seen in the Info that you gave me DOA 4.1.2.0, but may previous version was 4.1.2.2.

Am I wrong? What do you thing happened?

 

[Marco Kalter]

My guess is that the version info is wrong. Is the issue fixed?

 

[rguillen]

It seems that in almost of case it was fixed. But there is a particular case when it occur again. It's in a column NCHAR(2), it contain values from '10' to '99', but when it has 'MM' it bring 'M '.

 

[Marco Kalter]

This seems more like a display issue, where MM is too wide to display. Needless to say, there is no real difference between 99 and MM as far as data is concerned.