RE: Row level security

> This is what we want to do
> * One database
> * Most users (Companies) cannot view
>    each others information
> * Select users can view the entire database
> * Access to the files happens via all available
>   interfaces (i.e. not just green screen)

While the subject says 'row level security', the description sounds more
like you want to segregate Company A's data from Company B's users, etc.
Going on that assumption, the classic way to use OS/400 security to do that
is to make AUT(*EXCLUDE) on the current PF and create an LF for each
company.  Grant authorised users authority to the logical (company) they can
see, and OVRDBF after your selection to point to the 'right' one.

This presumes that user Buck doesn't need to see records from company A, B
and C at the same time, but he can see them one after the other, viz. Select
A, CHKOBJ, OVRDBF, see A.  Select B, see B.

Another way is to restrict all authority to the PF and create an API that
all code must call in order to read the data.  The API validates that the
user is able to access the record she is requesting.  That API can be
implemented as a stored procedure for the ODBC/SQL people, or as Scott
notes, a read trigger for V5R1.  That eliminates bulk transfers like FTP and
Client Access.  The good thing about stored procedures on our platform is
that the SP is an HLL program.  So the same program can serve green screen
or SQL clients.

  --buck