1 Reply Latest reply on Nov 9, 2018 8:46 AM by Joshua Milligan

    Prep - 11 howto's

    alex alex

      Hi all

       

      I tried Prep today, and I have several questions like "how to" I would like to find answers to:

       

      1. How to group several nodes?

      2. How to create a colored box behind nodes (in the background, not colored nodes, but just a colored box)?

      3. How to manually move/reposition the node?

      4. How to manually move/reposition the connector line between nodes (make additional "corners" etc.)?

      5. How to load several sheets (tabs) of excel file if you have column names present on 1st sheet only, and other sheets have same columns but without column name?

      6. If I have a table with several columns (in Excel source file), a cell of one of which has several text rows separated by "line break" - how to separate this text rows into several table rows duplicating other columns of this row at the same time?

      I need to change:

      Header 1Header 2Header 3
      One

      3

      4

      5

      Seven
      Two9

      Three


      Into:

       

      Header 1Header 2Header 3
      One

      3

      Seven
      One4Seven
      One5Seven
      Two9

      Three

       

      7. How to copy history items from one node to another?

      8. How to insert a line break inside a node comments box?

      9. How to increase the number of visible characters in a node name (before you get "...")?

      10. To have a control over completeness of the data: Is there any way to show totals / amount of records / other calculated fields/numbers   above the nodes or near nodes?

      11. What do you do to control completeness / correctness? If you have many nodes and you load new data which might be different from the one you tested your nodes on (the same column names but different contents) - it takes much time to go through every node to open it and see what happens there with the preview of result - what do you do to get a fast picture - like KPIs/checks on each stage of the flow.

       

      Thanks!

        • 1. Re: Prep - 11 howto's
          Joshua Milligan

          Hi Alex!

           

          Here are some thoughts:

           

          1 - 4: These are all not currently within the paradigm of Tableau Prep.  One highly requested feature is moving nodes (steps): https://community.tableau.com/ideas/8964 and you might want to add your support to it and additionally create new ideas for grouping or adjusting connecting lines.

           

          5. This one is possible, though not as seamless as it should be, in my opinion.  There is an idea for joining (and I would hope union too) based on position: https://community.tableau.com/ideas/8793.  A possible work-around is to adjust the Input Step for each sheet/file to not have column headers and then everything will come in as F1, F2, F3, etc...   You can then easily union based on column position.  You'll just have to rename the fields once the union is complete and filter out the headers that did come through on the sheet/file that had them.

           

          6. You should be able to split the field based on a line break -- try calculated fields like

          SPLIT([field], char(10)), 1)  
          and  

          SPLIT([field], char(10)), 2)

           

          Then you can pivot those into rows.

           

          7. Do you mean a Change within a Step to another Step? I agree this would be a very useful feature.  Please vote for it here: https://community.tableau.com/ideas/9133

           

          8. I've not tried, but try SHIFT + Enter.  I don't know if it's supported.

           

          9. Not currently possible that I know

           

          10. This would be a very useful feature.  Please vote for it here: https://community.tableau.com/ideas/9394

           

          11. This is actually a very good question and deserves more thought and consideration than I can add here.  I'd love to see some features included in Tableau Prep that give a better picture of lineage and even some smart automatic checking of data (the suggested data roles and fixes are a move in that direction - but a much more robust checking throughout would be awesome!)  Right now, I find myself checking my output against my input and accounting for logic - but that's not always easy, precise, fast, or in some cases even possible.

           

           

          These are some quick answers to your questions.  If any of your questions didn't get the complete answer you were hoping for, please start a new thread for that specific question so we can all go deeper into finding solutions!

           

          Best Regards,

          Joshua