Using CHAR semantics from an Oracle Database, Unicode databases for CHAR type columns (larger than 500 characters) and VARCHAR2 type columns (larger than 1000 characters) is not supported in this release.
The contents of tables that use more columns is archived, but restoring these entries to the reporting database is not supported.
Resolving the problem
A workaround can be attempted if you want to try to restore data; this workaround might not work in all situations. For example, this workaround cannot be used on the COMMLOG table.
1: Create an Insert request, and add a Local Column Map. When the columns are displayed, remove all VARCHAR or CHAR or NVARCHAR columns that are larger than 1000.
2: Include all primary key columns in the insert request.
3: Start the Insert request, in insert mode.
4: Add one of the columns that is greater than 1000 characters in
length. If you have any more column values to be inserted then perform
step 1 again with the new columns, and remove the previously included
non-primary key columns in the column map in the Insert/Update request.
5: How the Insert/Update request is processed depends on the number
of columns in the tables. Include the fewest numbers of columns in
each of the requests until all the data is restored to the COMMLOG
table. Subsequent Insert requests must be run in insert/update mode,
or the IBMtm Optimtm restore process fails.