Hi All - I am currently working on a project with what I believe is a rather common need: restrict users to their data (e.g. region) at a detail level, but some expressions reach outside of that set (e.g. Set Analysis.) An example of this would be seeing how their region compares against others at an aggregate level. I am curious if anyone else in the community has a good solution.
I have seen and done this in different ways in the past but always come across the same problems:
Using Section Access to reduce the data doesn't allow to compare against the entire set (I know this could be done with aggregate tables but I'm trying to avoid if possible)
Using Section Access to set "Initial Selection Based on Section Access" is a start, but the users can always select another value or clear (I've even gone so far as to lock and set a trigger to lock the field on unlock, but "Unlock and Clear All" will still clear.)
I've tried the "Always One Selected" on the Section Access field, but this causes problems when opening (the value selected on save takes precedence over the Section Access value)
Isaiah, The only reliable way to restrict unauthorized access to the data is data reduction based on section access. It means that the data outside of the avaiable region simply doesn't exist, and the only solution is to use aggregated tables available for all users. (I'd be glad to learn about an alternative way.)
Thanks Michael - I figured that would be the case 😃
I have had good luck in the past in a QVS-only environment by disabling collaboration (e.g. sheet objects) and using a combination of variables and flags to restrict what a given individual or group can see. Certainly this is more of an obscurity model than security.
I´ve currently exact the same requirement in a project. (rep should see his area in detail, but should also see for benchmarking aggregateds value for higher levels, e.g. region, country).
Section access will cut data of higher levels and I will loose benchmark values of higher levels.
At the moment I only see the possibility to pre-aggregate data for higher levels with Scripting and put it in the data model as different fact type.
What was your solution to solve this problem?
Are there possibly new ways how to solve it in current version of QV?
