Collation Specification at the Column Level
This feature allows the specification of a collation sequence at the column level that differs from the database default collation sequence. A new optional COLLATE clause is added to the column specification on the CREATE TABLE statement. The COLLATE clause lets you specify a case-insensitive collation for columns that contain Unicode data.
In previous releases of Ingres, the DBA had the option when creating a database of defining a collation sequence to be used for non-Unicode text columns. The Unicode standard default collation sequence was provided for Unicode text columns. These collation sequences were in effect for all columns in the database and could not be changed without recreating the database.
Note: As part of this feature, the data descriptor used throughout the Ingres system changed. This data descriptor is also compiled into imaged ABF applications. After upgrading to Ingres 2006, all ABF applications should be re-imaged. Delete the contents of the ABF object directory, $ING_ABFDIR/database-name/app-name, and then re-image.
For details on this new feature, see the SQL Reference Guide.