5 Replies Latest reply on Jun 11, 2011 4:30 PM by tommyodell

    Dual axis question

    R R


      a while ago I posted a question to get a combined bar/line chart. The problem is, that there is a nested structure in the x-dimension (multiple values per refinement level, see picture 1). I got some answers, but they are rather hacks.

      I tried to re-do some, but still stuck with the shown result (see picture 2). The line won't be calculated globally, but just locally. So close, but no idea how to solve that in order to get a single line (as in picture 1).


      (1) Any suggestions?

      (2) There is an interesting alignment of the measures in the ROW shelf (picture 1) similar to the one that one gets if using a second axis. This one aligns the measures vertically. Any idea what this means and how to get that?


      Thanks in advance,


        • 1. Re: Dual axis question
          Joe Mako

          Here is that previous thread:



          For your attached images, the difference is in p1, you have a continuous measure (a green pill) on the Columns shelf, and in p2, you have two discrete dimensions (blue pills) on the Column shelf. The discrete dimensions are categorical, and when there are more than one on a shelf, Tableau will create a Small Multiple display, creating panes, and in Tableau, lines cannot cross panes.


          The main reason my previous workbook provided feels like a hack is because Tableau does not support clusters of bars, if you are okay with the spacing between all bars to be equal, then I think it feels less like a hack. Attached is an example without the clustering.

          • 2. Re: Dual axis question
            R R

            Thanks, Joe!

            Your "hack" helped me a lot and I didn't want to complain. I just thought there might be a way to integrate the line in a more "convenient" way. This seems not to be possible to the (possibly meaningful) constraints of Tableau. Fine for me as long I can reproduce it. What did you use to reshape the data? Data wrangler?


            Also can you throw some light on the second question? How can I get such a alignment of the measures? What does it mean?


            Thanks again, Joe!


            • 3. Re: Dual axis question
              Joe Mako

              >  What did you use to reshape the data?

              Whatever tool feels like a good fit. For data you can fit in your clipboard, Data wrangler is a great option, and for data that is large, I am currently enjoying http://kettle.pentaho.com/


              As for the pills on the Rows shelf, the shelf is sizable, so if you hover your mouse pointer on a edge, you can grow or shrink it, useful when you have a bunch of pills.

              • 4. Re: Dual axis question
                R R

                Thanks again! Didn't know about the resize shelf option ...


                • 5. Re: Dual axis question

                  On the subject of data reshaping... thanks for the links Joe. I've heard DataWrangler mentioned so many times, and now I've finally taken a look. I'm not familiar with the Pentaho Data Integrator, but I'm downloading to give it a whirl. Until now I've done all of my data reshaping with the (very awesome) 'reshape2' R package.


                  If feel that data reshaping would be an incredible addtion to Tableau.


                  There are ways to get around Tableau's lack of support for this, but the ease with which I can get data reshaped and into Tableau depends quite a bit on whether it's a "Long to Wide" or "Wide to Long" transformation, whether or not the dashboard needs to connect 'live' to a database, or whether a flat file is okay. I can always reshape the data like I want in a flat file using R/reshape2, but that doesn't help if my dashboard needs to be 'live'. With 'live' DB connections, it seems that some RDBM systems have some reshaping functionality, but it isn't universal. And unfortunately the ones that I work with don't support it.


                  It would be amazing if Tableau could support reshaping in-app, without having to rely on the SQL queries.