2 Replies Latest reply on Jan 29, 2014 11:20 AM by Jennifer Hill

    Can I create a user filter based on blended data?

    Jennifer Hill

      I'm pretty familiar with the concept and execution of user filters.  I've used them successfully in several projects, and they generally take the form:

      username () = local_identifier


      This works fine when there's one identifier for each data row.  However, I need a new type of user filter which gives individual end users access to large subsets of data, with which their individual identifiers may or may not be affiliated.


      I uploaded a supplemental Excel file, which lists each local identifier and all of the department groups to whose data the individual end user must have access.  For example, Bob Smith should have access to his own sales data per the calculated field username () = local_identifier.  I also want Bob to have access to all sales data for states NY, PA, CT, MA, ME, and NJ.  By extension, Bob's colleague Tom Brown requires access to his own sales data, and also all sales data for NY, PA, and FL.  Notice that the groups are not mutually exclusive, and there may be overlapping membership.  The supplemental table I created looks like this:


      Bob     NY

      Bob     PA

      Bob     CT

      Bob     MA

      Tom     NY

      Tom     FL


      I was able to blend the data on a common dimension (our local identifier).  My question is:  how do I create a new username field that is based on group memberships/access defined in the secondary data file?  I suspect it would look something like this:

      if primary.username () = ATTR(secondary.local_identifier) then "Show" End

      This calculation is valid (yay!), but it seems stuck as a Measure. I know this has something to do with the ATTR function, but I can't make the calculated field work any other way.  I can't drag it up to Dimensions, and I can't drag it up to the Filter shelf.  My end goal is to filter the data on this user filter.


      I'm not able to share a copy of this workbook due to data confidentiality.  THANKS FOR YOUR HELP!