Jonathan, I looked at your workbook tried to figure out a solution, unfortunately without success. You answer so many questions here on the forums, I just wanted you to know that someone was giving your question some attention. Maybe tonight Richard or one of the other folks will have a go. (This will at least bump it to the top for you.) Sorry I couldn't help.
Thanks for looking at it!
3 of 3 people found this helpful
Got help from Joe Mako answering this one, here are the details:
1. Change the date fields used as dimensions on the Columns shelf into a single continuous date, for example with a calculated field like:
Put this new field on the Columns shelf and set it to Exact Date/Continuous.
Why: The Tableau datepart dates (i.e. month([date]) = 7) are discrete (blue) dimensions, and even though the new datetrunc dates in 7.0 (i.e. month([date]) = 07/2012) show up as continuous (green) pills, they are still treated as a dimension and cause table partitioning. The calculated field is truly continuous and does not cause partitioning.
2. Move the Month([Date]) from the Level of Detail onto the Path shelf.
Why: On the Level of Detail shelf, the Month([Date]) causes partitioning to happen. On the Path shelf, it is used to set the order in which Tableau will connect the lines.
3. For each measure, set Format->Hide (Connect Lines).
Why: This is the secret sauce that enables Tableau to get past the null values coming from padding and/or underlying data.
Nice! Thanks for this Jonathan.