I am writing a jdbc driver with only publisher channel to load in eDir
an ORACLE table/view.

This source table have n rows for each users:
user's key is duplicate for each profiles.
The record of the table have the follow structure: <user_id +
profile_id + ... + other column>.

I have created a custom multi-valued eDir attribute on User class and I
have mapped this attribute with the column that represent a list of one
or more profiles (the relation user-profile is 1:n).
The primary key of the table is composed of (user_id, profile_id).
The Matching rule is based on user_id and the problem is that for each
rows (for each profiles of one user) the driver association is just
present after the first row.
So I remove association after the operation "do-add-dest-attr-value",
but does not work because the flow of records in reading is not
sequential and the driver reads more than one record at a time.
Alse I don't understand what kind of operation I have in input: add -
modify - delete?

Someone had the same problem and can help me?
Thanks in advance.

savoialuisa's Profile: http://forums.novell.com/member.php?userid=58872
View this thread: http://forums.novell.com/showthread.php?t=379939