1 Reply Latest reply on Oct 7, 2016 3:00 PM by Patrick Van Der Hyde

    Cartesian Join

    tommy durant

      So, i want to create a heat map to see if our customer spend by product.  The heat map has customers on the rows pill and products on the columns pill with revenue on the color pill and detail pill.


      If they do not buy that product, its not in the underlying dataset.  so, Tableau simply has a blank under that product spot for a specific customer in the heat map.


      my end user request is that it must have a shape/color for missing products.  I can get this to work but the issue is i have to use a cartesian join to get every customer and product combination. then i join in revenue.  This gives me a row with no revenue or 0 revenue for a customer who does not purchase that product.  I do all this in my underlying sql server database.


      If i exclude the date fields from this join and just do an overall sum of revenue for every customer and product combination, this is 355,000 rows.   if i add in a month field,we are starting to talk about 10's of millions of rows.  My underlying sql database has issues with this. and im unsure how slow the dashboard would go under this situation.


      Is there a better way to go about this?   It would be nice to have this done at a month level , so i can have time trends to this view.