2 Replies Latest reply on Feb 6, 2012 7:56 AM by Alex Kerin

    Data blending - same fields, different observed distinct values

      I am trying to create a stacked bar graph where the column is a time dimension and the row is a sales measure.  Each observation within sales has a product description.

       

      I have two databases that use the same tables and fields (although observations will be different; also, this setup was not my design, but it's what I have to work with).

       

      Observations from table 'sales' in data source 1 (set as my primary source) realize the following values for the 'product_desc' field - {A, B, C, D, E}.

      Observations from table 'sales' in data source 2 realize the following values for the 'product_desc' field - {B, C, D}.

      So the set of distinct values from data source 1 is a superset of those from data source 2.

       

      My visualization currently is a bar chart of sales over time.

       

      When I attempt to add the 'product_desc' dimension as a 'level of detail' (specifically in color so that I have colorful, stacked bars), the result is not what I expect.  The bars increase in size (double counting?) rather than Tableau partitioning each bar by 'product_desc'.

       

      Any suggestions?  Any help would be greatly appreciated.