On reflection I'm aware a little bit more info is needed.
An example table structure would be:
Leader Staff FTE Salary USERA USER1 1 15000 USERA USER2 1 18000 USERA USER3 0.5 12000 USERB USER4 1 26000 USERB USER5 0.8 23000
Both USERA and USERB need to see all FTE information but should only be able to see the Salary info where they are listed as the Leader. Their Qlik login matches the Leader field.
The problem is that you want to display all data from some fields and at the same time reducing the records from other fields and this is conflicting.
Some years ago we had had a similar case and we developed a very complex solution with huge loops to create the section access tables, which was also loaded twice and needed some variables and triggers and also the requirement that a user couldn't change objects or create own ones - it worked but it was a maintenance nightmare ...
Therefore we replaced it after some time with two specialized applications which both fulfill their requirements much better as our first matching of too different applications. This might be the easiest way ...
If this is not possible or preferable you should give this a try: Mask or de-identify data for certain users using Section Access.
I think I have a solution for you, where you declare a field in Section Access, then use that field in Section Application to limit access.
The setup might be something like:
Load ACCESS, USERID, UPPER(USERID) as LEADER, OMIT
From [your user list];
LOAD LEADER, STAFF_ANYONE, STAFF_LEADERONLY;
SQL SELECT u.USERID AS LEADER, d.Staff as STAFF_ANYONE, case when u.USERID = d.Leader then d.Staff else NULL END as STAFF_LEADERONLY
from [your user list] u
cross join (select distinct Leader, Staff from [your leader-staff table]) d;
Load Staff as STAFF_LEADERONLY, Salary
From [your salary data];
Load Staff as STAFF_ANYONE, FTE
From [your salary data];
The result of all this is that each LEADER can see anyone's FTE data, but only his/her employees' Salary data. It's all in the way the tables link together, starting with the LEADER field in the hidden table in section access. That links to LEADER in tblReduction, and then the two STAFF fields link from there to the two data tables.
My underlying data reside in SQL Server, so I use CROSS JOIN to get the cartesian product of the two tables (so that every Leader can see every employee's FTE data). In Qlik, you can achieve the same with OUTER JOIN.
If I understand your suggestion right does it mean to develop two independent datamodels within the application because you will for example need different fields for the EmployeeID and also to link these data to further dimension-tables like a calendar. If you connect them in any way you will apply the section access data-reduction again and/or creating some synthetic key or circular loops.
In some cases two datamodels might be a practically way but for many scenarios it will be rather not suitable.