<< Back

New in Tableau 10.2: Join Calculation

Note: Tableau 10.2 hasn’t officially released yet and this video is using the Beta version (Pre-release). You can sign up for the beta here.

In Tableau 10.2 we’ll see a couple of great new features. You can see a whole list here; which includes new ways of formatting lines, using spatial files as data sources, direct share point integration and many more.

In the video below I use a new feature called the Join Calculation. I use this new feature to tackle a very common issue I see many of our clients encounter. A dataset that contains columns with a start and end date, of for example, mobile phone contracts or when people are admitted to hospital. The problem here is that this is difficult to visualise over time. In the video below I describe how you can create a product join (e.g. a scaffold) in order to join a data set with dates by using the new Join Calculation feature.

With a Join Calculation you can now create custom joins through calculations which mean you can bring multiple data sources together, even if the field names don’t match.

New in Tableau 10.2: Join Calculations

Andre de Vries

London, UK

9 thoughts on “New in Tableau 10.2: Join Calculation

  1. This is terrifically helpful. I’ve been using Excel legacy files and doing greater than / less than joins with dates and minutes through legacy connections, but was fearful that once legacy versions of Excel stopped being supported, I would have no alternative. Even before 10.2 comes out, your video showed me that all I have to do is add an ID field with “1” in each line of my original database, then link it to the ID field of the time database, and then create the greater than / less than Boolean calculation using a calculated field.

    I tried this with a small test database. Now I’m wondering whether this approach will speed things up with a larger database. I often work with a database of 3000 records, and I visualization start and stop times during a day, which means my time database has 1440 records (12:00am to 11:30pm). 3000 x 1440 is a very big number, and visualizations through my legacy approach take a long time. I’m hoping that using calculated fields, the time will shorten somewhat.

    But in any event, terrific video — one of the most helpful things I’ve encountered in Tableau in a very long time.

    1. Hi Andre,
      I have two custom sql queries in two different sheets. Both the custom sql queries come from the same data source. I cannot join the two queries because of many to many relationships. So here is the issue, I have one table (table 1) that has year starting from current year July to next year June calling it a Fiscal year. So, the one query which already has the fiscal year is good, and I can pull employee’s salary paid within that fiscal year. The other table, which is a little corky, is the table ( table 2), which has the salary changed as on an effective date. Meaning, if there has been any change in the salary due to fulltime/parttime or change of position, there is a change in the salary and only that effective date is captured. I sorted out a couple of issues and got the two custom queries ready. So I am trying to join the Fiscal year from one table( table 1) to the Fiscal year which is a calculated field based on the effective dates in the other table (table 2) .
      Now I am creating a dashboard and I need a single option to click within the filters and when I do, I should have data for that one fiscal year. I have done that for employee ID that is common ID in both tables, which came from the tables within the custom sql queries by using Data blending and editing relationships. I am unable to join the Fiscal year and I don’t know why . Do you have a solution ? I am working on Tableau 10.5


  2. Hi Andre,

    great video! The ability to join on calculated fields is indeed something I’m also loving a lot in the 10.2 beta.

    That said, I think the use case you showed in the tutorial can be solved a lot easier using Gantt charts and a simple calculation. This workflow, which I’ve quickly recorded (see https://www.youtube.com/watch?v=FxHDR4vm-04), has three advantages over the one you presented here:
    1) It’s not necessary to create the additional date table.
    2) Since there’s no date table involved the method is not sensitive to data points outside the range of the date scaffolding table (e.g. in 2012).
    3) It only generates one mark per customer – in this case 8 as opposed to the 6,227 in your example. This might especially become relevant when working with more data.

    Cheers for the inspiration and happy holidays!

    // Konstantin

  3. Nice work. I previously used Alteryx to join the data so that I could do the same type of analysis for account receivable balances. Accounting programs will provide invoice data and payment date so you run into the same situation as your phone contract example when visualizing balances over time.

  4. very nice.. Great easy example to follow. whether it’s relevant to the problem is a moot point – because It left me understanding how I could apply this technique elsewhere, which I’m sure is what you wanted.. 🙂

    I have a feeling this is going to help me with an intractable problem on v9.2 trying to create sankey diagrams and pad out the necessary data on the fly…

Leave a Reply

Your email address will not be published. Required fields are marked *