Some data model may yield weird results when filters are applied in certain fashions. For example, I created a data model whose date filter can only work properly in “last N days” fashion since it requires that one edge of the time frame is “today”. If I could be able to limit the filtering options available for the user to the supported ones, it would be very nice. Now I’m compelled to warn the user with a note whose efficacy is very limited and anyway doesn’t prevent the user from forcing a setting that would return wrong results.
This is indeed valuable. Though, could you share some real-world examples of what you mean by “one edge of the time frame is today”?
Yes, sure.
I have a table of members of an on-line community refreshed every day at midnight: one attribute is “last_login_date” and, unfortunately, a detailed activity log is not available.
With the data I have, I can easily expose data such as “number of users active in last N days” simply using “last_login_date” as the field for the last N days filter. Any other option for filtering not including today as an edge (actually yesterday as a whole), such as between or even last N month would yield wrong results.
However, this is just one case. I faced many in which limiting available options would be required to guarantee correct results for users.