3 Replies Latest reply on Sep 13, 2017 10:32 PM by Shaik H

    Tableau Data Extract Refresh

    kon kon

      Hi Gurus, newbie tableau site admin here. One of my tasks is to productionalize the tableau workbooks created by developers. Here is what i usually do:

      Step1: Download the packaged workbooks from DEV Server/site, then create a local tableau extract using an Application ID and then publish to PROD(Using Tableau Desktop)

      Step 2: Schedule the tableau extract on PROD server, then manage security etc.

       

      Recently i am running into an issue in my Step 1 when creating my local extract. Usually when generating the extract i see the a couple sub-steps

      After downloading workbook i have to re-enter credentials >  Takes me to Data Source Tab > Make sure Extract is selected > Click on Update data > Click on any worksheet

      Step 1a)  Creating Extract > Import data - Rows retrieved . Here i can see the rows being retrieved from my oracle table.

      Step 1b) Creating Extract > Executing Query

      Then it takes me to the worksheet.

      But recently, 2 new columns were added to oracle table and so they were added into the Workbook Datasource. Since then when i download the workbook/create the extract i am seeing

      that the first step 1a) has changed.  I see another step that is taking a long time. It just says

       

      Creating extract database

      After this is complete i am seeing the Step 1a) Importing data > Rows Retrieved step. Any ideas why this new 'creating extract database' is occuring now? Eventually the data refresh is getting completed but it almost taking double the original time to refresh( and it doesnt make sense as i just added 2 new simple dimension columns). Any thoughts are appreciated.

       

      Environment: tableau desktop 10.2, tableau server 10.2

       

      Thanks,

      Kon

        • 1. Re: Tableau Data Extract Refresh
          Toby Erkson

          I doubt this is a Tableau Server issue.  Adding the two columns could easily be accountable for the increased time.  I recommend making sure the workbook is optimized.  Here is a place to start:  Optimize Workbook Performance

          • 2. Re: Tableau Data Extract Refresh
            Alan Perez

            friend, it is because the extract is being updated with the new columns and spor it tells you that it is creating the extract again, but if you look now it does not give you the option to save the extract, since that extract is only doing refresh

            • 3. Re: Tableau Data Extract Refresh
              Shaik H

              Hi Kon,

               

              This might help you.

              Setting Different Refresh Schedules for Each Extract in a Workbook | Tableau Software

               

              Just an idea for a manual workaround, that may not be helpful in your situation:

               

              0) You start with your original extract, which you should have in a twb file, so that you can switch back to your original live connection at any time.

              1) You have your extract saved separately as a tde file, and open it directly from Tableau to generate a new extract based off your original, with filter leaving out the dates for which you want your data refreshed. Save this new extract as a separate tde file, you will be adding rows to it later - essentially this contains only the "old" data, that will be left intact.

              2) Open your original twb file (from step 0), switch to live connection and generate a "new" extract filtering ONLY to the dates containing the new data - this is what you need to add to the file from step 1.

              3) Open your extract containing the "old" data in Tableau. Although it is an extract, you will have to tell Tableau to make an extract out of it once again to activate the "Extract" submenu in the File menu.

              4) Go to Extract->Append Data From File... and point it to your extract from step 2 containing the "new" data only.

               

              You end up with an extract that contains your "old" and "new" data.

               

              Not sure if this helps at all though, as:

              1) It is entirely manual process, very prone to errors. Probably some of the gurus could help you figure out if you can automate this via tabcmd, but to the best of my knowledge, there is no way.

              2) You are creating your "old" extract out of extract, but still essentially this wrangles the entire available information, so I do not know if it will be faster than referring to the original datasource to make a full refresh.

               

              Good luck!