unicode stores multi-byte characters, the non-unicode stores ascii 7-bit characters. The characters that you see as ?? are probably 8-bit ascii characters.
Remedy uses the NLS_LANG and LANG environment variables on the server. On our unicode instance we use american_america.AL32UTF8 There are other environment variables that drive certain things in Remedy that are not directly related to Oracle: LANG - for unicode db we use en_US.UTF-8 LC_ALL - for unicode db we use en_US.UTF-8 Unfortunately, the 8-bit representation of a character on a non-unicode db does not match the representation on a unicode db. That is why you see a character on one server, and questions on the other. On your non-unicode arserver, the NLS_LANG is probably set to US7ASCII or WE8ISO8859P1 or something similar. Your unicode db is probably set to american_america.AL32UTF8 or something similar. Unfortunately, the two are not interchangeable due to the difference in character sets of the 2 oracle instances. You have to remember that everything is abstracted through the remedy api when it is passed through DSO. Some good references: http://www.adp-gmbh.ch/ora/misc/globalization.html http://www.oracle.com/technology/tech/globalization/htdocs/nls_lang%20faq.htm Unless you use these characters in workflow names, view names, form names, etc. you probably won't have a problem. Once you have both of your instances using the same code page, you can perform some cleanup to convert the dirty data to something more meaningful (e.g., ' to ' - yes, they are two different characters) Not an expert in that area, but that's my take. Axton Grams On 4/20/07, Sabrina Ethridge <[EMAIL PROTECTED]> wrote:
** I've been going back and forth between support and our DBAs on this problem and have gotten stuck, so I need your help, listers. We are using DSO to transfer data between two ARS servers. The source uses a Oracle 9i(non-unicode) database and the destination is an Oracle 10g (unicode) database. Both are on Unix Solaris 5.8, ARS 6.3, patch 20. Some records have control characters in fields, such as a tab or a return character, from an end-user cut and pasting into a text field from another application. When these records are viewed via the WUT from the source(9i db), the control characters do not display anything but whitespace. When viewed via the WUT from the destination (10g db), the control characters display as question marks (??). Our DBA is convinced that this is not an issue between unicode and non-unicode, since when she uses SQL*Plus to look at the records directly in the database, she sees question marks instead of whitespace in the records in both databases. Tech support, however, says that the database charactersets must match. Our DBA is convinced there is a LANG parameter(or something) that tells Remedy how to talk to the database and that we have these mismatched on the source and destination servers. Does any one know if there are LANG parameters that Remedy uses to talk to the database? Or any ideas on what could be wrong? TIA Sabrina Ethridge [EMAIL PROTECTED] ________________________________ Ahhh...imagining that irresistible "new car" smell? Check out new cars at Yahoo! Autos. __20060125_______________________This posting was submitted with HTML in it___
_______________________________________________________________________________ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org ARSlist:"Where the Answers Are"
