8.10/8.95_F1 Upgrade - Unicode Conversion for Business Data/Control Tables

List,

Upon completing initial upgrade of 8.0 to 8.10 system (8.95_F1 Tools), I noticed that the database size for upgrade testing (PS_DEV) has extended from 91GB to 126GB on MS SQL 2000 Enterprise. I was preparing to execute the Unicode conversion for this database, but upon examination, it appears that the conversion has already occurred. In every instance, the CHAR data type has been converted to NCHAR, etc., and the extension of the database is appropriate for Unicode conversion. It is important to note that within the Upgrade Plan, we selected the "Unicode" option within the Business Data and Control Tables segments of the workbench (under Advanced), and the database data sources for this item currently appears with Unicode activated. Did this selection during the upgrade workbench actually execute the Unicode conversion? I have read at length that this conversion must be done after the fact using the Unicode Conversion workbench, but I cannot see any other explanation for this. Can anyone clarify this or shed more light on the process?

Thanks in advance.

Yes, if the data sources were set with the Unicode flag, then most likely it did. While troubleshooting our Unicode conversion issues, the support person we were working with actually mentioned creating data sources with the Unicode flag set, and then running R98403, which would in turn create the tables (with Unicode data types), copy the data over and perform the conversion, all in one fell swoop. We didn't want to go this route as it would have taken much longer than running the coversion using the native DB processes.

Affirmative.

GSS is saying that selecting the "Unicode" option during the upgrade plan execution is neither advised nor supported. They are saying that if you do choose it, it does execute a Unicode conversion on the fly of only the tables that are touched by automated table conversions, but not others. If you run the post-upgrade Unicode conversion, it skips these tables and converts only the tables that are in non-Unicode format So, essentially the answer is to leave this option alone for the Production table conversions, then execute the Unicode conversion by itself once the TCs are complete.

No answer as to why this option is even available within the workbench. They seemed shocked that the data even worked.

Thank you for the response.