3 Replies Latest reply on May 24, 2018 7:42 AM by mark santino

    Stopping Tasks / Not Autorunning Flows?

    Jason Turner

      Hello,

       

      I'm working with some large text files (client required provision, essentially large CSVs that we need to make tall). Based on a few initial trials, we're looking at ~1GB .hyper extract created from steps in Tableau Prep (pivots, mostly).

       

      I imagine using smaller samples will "fix" some of the runtime issues, but overall, the inability to quit a flow calculation or stop flows from reloading while attempting to do several tasks simultaneously forces me to ctrl-alt-delete several times. In essence if you jump between data sets quickly while doing repetitive tasks like column renames, you can very quickly overwhelm Prep and cause it to run the flow for quite a long time (stopping other operations) with no ability to stop it.

       

      Is there an option to NOT autorun flows (e.g. just allowing me to manipulate column headers, etc., before recalculating the finished data set all at once?)

        • 1. Re: Stopping Tasks / Not Autorunning Flows?
          mark santino

          Hi Jason -

           

          At the moment the ability to halt a flow at any given point does not exist. In regard to your last comment about ..."recalculating the finished data all at once", any changes made in a Clean step (i.e. column manipulation, grouping, etc.) is conducted on the sample level NOT on the full scope of data. The default sample might be a certain # of rows off the top - limited in size, and faster to process, but probably not very representative of a large data set; a random sample will be more comprehensive. Either way, any action taken prior to writing to an output is conducted on the sample set. Output files contain ALL the data.

           

          Regards,

          Mark .....

          • 2. Re: Stopping Tasks / Not Autorunning Flows?
            Jason Turner

            Mark -

             

            Thanks for confirming my understanding of what was probably causing the issue.

            Either way, any action taken prior to writing to an output is conducted on the sample set. Output files contain ALL the data.

            Be that as it may, even on smaller problems, I've run into numerous times where I've been forced to quit Prep due to seemingly infinite flow run times on sample data (Which prevents doing stuff like column renames or even saving).

             

            Specifically, it almost seems like requesting flows run on multiple data imports (In my case .CSVs) never resolves.  I've since set my "sample" rows down to 100, which has VASTLY improved my ability to avoid lockups, since it's calculating on trivially small data sets at that point.

             

            I strongly suggest the ability to *not* run the flow at all, and, honestly, NOT have the default setting be to run the entire flow automatically. (Or, alternatively, to give the user a method to stop calculation).  While the statistical summaries shown in the windows are very cool, and could be very useful, they're only useful if they're accessible.

             

            Secondarily, it may be wise to consider the setup data workflow tools like Alteryx or KNIME have, which visually indicate % completion and flow status throughout the workflow.

             

            As-is, Prep has become very frustrating when trying to union several data sets, and the "large" (ish) data sets that I'm using should be the anticipated use case for the software.

            • 3. Re: Stopping Tasks / Not Autorunning Flows?
              mark santino

              Hi Jason -

               

              Did you participate in the beta for Tableau Prep? It was quite comprehensive, and solicited a ton of feedback. If you did participate, you're probably on the contact list for update notifications on the the product. Regardless, as an FYI, the upgrade schedule for Prep is an aggressive one, and totally independent from Tableau Desktop and Tableau Server. I honestly don't know if the items you mentioned are being rolled-in anytime soon, but I think you can expect lots of changes to the product in a relatively short time period.

               

              Regards,

              Mark ....

              1 of 1 people found this helpful