1 Reply Latest reply on Feb 14, 2016 5:28 AM by Yuriy Fal

    Territories and zipcodes - User Filter


      Is there an efficient way of creating a user filter to allow 70+ users to see zipcode-centric data? Each user would only be allowed to see his/her zipcodes.



        • 1. Re: Territories and zipcodes - User Filter
          Yuriy Fal

          Hi cignior,


          1) A User Filter is essentially a static Set,

          so it is created (and further maintained) manually,

          so any changes to it should be done manually, too.


          2) As they're Sets, there may be more than one User Filter per datasource.

          Having more than one User Filter per DS, one could implement

          a "hierarchy" of User Filters -- for example by mapping "geo" fields

          (Country / State / County) with specific User Groups from TS / AD --

          and manage them by including Users into (several) User Groups.


          3) A User Filter could be just copied from one datasource to another --

          if a common "base" field does exist (same name and -- important -- same values,

          though they don't be complete in each DS, just an overlap between them).


          Better still, UF could be defined as a Datasource Filter in every published datasource,

          so ordinary Tableau users can not change it, even if they're in Web Edit mode on TS.


          4) A reasonable suggestion is to have a local "UF Master Datasource" workbook

          with every field used for every User Filter available elsewhere.

          It could contain a number of simple "One-field Dimension table" datasources,

          one for each User Filter. Or there may be a so-called "Scaffold" datasource,

          which is a cartesian of all dimensions used for User Filters.

          Either way, it is worth the effort to have a complete domain

          (a number of unique values) in each of those dimension fields.


          One could define, change and test User Filters here,

          and then copy them to their "destination" datasources / workbooks.


          5) Since one could define as many [User] -- [Dimension] relations

          as needed (actually by adding rows to the User Filter Set),

          a User Filter is a potentially "many-to-many" kind of Set.


          It is similar to a so-called "bridge" table in a database schema,

          but with a distinction, that a bridge table typically (inner-)joined 

          between fact and dimension tables, and could change the granularity

          (duplicating facts because of many-to-many relations).

          User Filter is a filter, so it doesn't change a granularity of a datasource.


          Hope this could help to understand User Filters

          and find a convenient way of applying them to your purposes.