3 Replies Latest reply on Aug 29, 2018 1:13 PM by Ken Flerlage

    Join Expectations

    Scott Carman

      Morning All,

       

      Ok so my current project is working on contract information.  I am trying to join the data as it is faster.  I am open to a blend but not sure how to do that and make it work with a hierarchy involved so that we can filter data based upon the users level in the organization.

       

      Here's the set up:

           Table 1 has the contract details such as the period of performance (Start and End Dates), contract officer, version number, etc.

           Table 2 has the financial information such as the obligations and commitments.

       

      One caveat is that this is a many to many relationship.  Sometimes the version number will change yet the PoP won't change.    Sometimes the PoP will change and the rest of information will stay the same.  Does  not a distinct pattern  The way I have overcome this is by using a series of LOD expressions which works.

       

      I have been trying a left join with a subset of the data.  Table 1 has 238K rows and Table 2 has 358K rows.  My return is about 1,048,000 rows.

       

      My issue is this - does this meet my expectations?  Not sure as I am not totally accustomed to working with such a large dataset. and the full datasets are HUGE!  A couple million rows a piece. (I have talked with our DBA and they refuse to join the data prior to Tableau.)

       

      That said, sometimes when I try to create the joining both Desktop and Prep stall out on me and just run in an endless loop.  Like I said I am open to a blending but I have run into filtering issues when trying to filter data based upon a hierarchy.

       

      Any suggestion or advice on whether my expectations for a join are met or not?

       

      Thanks,

       

      Scott

        • 1. Re: Join Expectations
          Ken Flerlage

          Unfortunately, I think we'd need to know a bit more about the data to properly answer this question. Are you able to share an anonymized sample of the data?

          • 2. Re: Join Expectations
            Scott Carman

            Ken,

             

            Unfortunately, I am under an NDA to not share the data.

             

            So I am starting to understand how challenging a question like this is to answer for someone who can't look at the data.  I have tried to ask this question in a variety of ways without success. That said, I sat down last week with Tableau Prep, which I got on Friday last week and was able to work out.  The way that it shows the join with what data is included and not, matched and unmatched, really helped me help myself.

             

            Thanks again for at least trying to help.  That's what I love about this community, everyone is so willing to help even with the abstract ideas!

            • 3. Re: Join Expectations
              Ken Flerlage

              So glad you were able to figure it out. Be sure to mark you last response as the correct answer so people can quickly see that, in this case, Tableau Prep allowed you to correct the problem. That will also allow us to close the thread.