RLS: with a custom postgresql query

I am trying to put row level security on a Custom SQL dataset that works perfectly as author. I have created a GroupName column and created, in the datasets I want to share a column with such name.

Everything seems to go fine

When I try to apply the filters (Apply datasets), QUickSight gives me a messagge saying “Share failed” and then "Dataset rules contain column(s) with unsupported data types: [ id_modello_test, id_risposta, id_domanda, id_risposta_prev, id_sondaggio, id_somministrazione, punteggio, peso, id_modello_test_vr, punteggio_risposta, flag_non_applicabile, numero_allegati, id_argomento, anno, id_argomento_father ]. Error code: DatasetRulesInvalidColType.

What is strange is that the listed columns are perfectly managed by the business intelligence.


Can you clarify the last line of your statement where you indicated that “the listed columns are perfectly managed by the business intelligence”. What did you mean by “business intelligence”? There is documentation here with respect to supported data types. Can you confirm if all the fields you are accessing are of the supported types? Supported data types from other data sources - Amazon QuickSight. Also, please see the following documentation on rule creation in the event that something there applies. Using row-level security (RLS) with user-based rules to restrict access to a dataset - Amazon QuickSight.

If everything above is good, if you could share with us any additional information about your progress we will try to help.


Thank you Sean! With the expression “the listed columns are perfectly managed by the business intelligence” I mean that I have already made dashboard and analysis on such data an they work perfectly well (so there is no problem of supported data types). BuT when I try to apply RLS I get the error I tell you. Thank you

Thanks again Sean. I think I have solved everything. The solution was just to read better the guide on RLS (although I think it’s not written very clearly in my opinion).