2 Replies Latest reply on May 20, 2011 7:53 AM by guest contributor

    Publishing extract datasource

    . bjorns

      Hello,

       

      I'm really fresh on Tableau. And am wondering about publishing extract data sources.

       

      Since using Tableau data extract is the recommended way of accessing data for performance, we do naturally want to use that for our reports. Our aim is to refresh extracts from an SQL server on hourly,daily or weekly basis... Then have whatever workbooks that are created import to those extracts.

       

      Workbooks would be created and viewed by many different users but share the same extract data sources.

       

      The problem is that the published extract data sources all reference the extract files on the local computer where the workbooks were created. So when someone else is trying to use your published extract data source on the server it fails with a file not found error.

       

      After reading around on the forums a bit more I'm starting to suspect that Tableau wasn't designed to do it that way. But I have a hard time wrapping my head around it. What would the purpose be to publish an extract to the server if only person can use it?

       

      Now I'm trying to figure out a way around that problem. My first thought was to use a shared drive on the Tableau server to put the extracts on. But it seems like the desktop client doesn't deal with UNC paths. This would mean that everyone accessing the extracts would have to use the same drive letter. Doable, but it feels clumsy.

       

      How us everybody else going about this?

       

      Every 5 cents welcome!

       

      /Bjorn

        • 1. Re: Publishing extract datasource
          Robert Morton

          Hi Bjorn,

           

          Tableau does not currently support a mechanism for multiple workbooks to share the same extract in some centralized location.  Several customers have requested this feature for Server, so it is on our radar.

           

          But let's take a step back to address your assertion that Tableau Data Engine extracts are the best approach for performance.  If you have a fast database system, you should absolutely plan to take advantage of Tableau's ability to issue live queries to your centralized database.  This "direct connect and go!" approach works very well for databases such as SQL Server.  Do you find that direct connections do not give you the performance you desire?

           

          -Robert

          • 2. Re: Publishing extract datasource
            guest contributor

            I agree that it would be very useful to be able to store an extract in a centralized place where multiple users can access it.  We are using SQL Server and the "direct connect and go!" option currently and the queries take a couple minutes a piece to load (each worksheet).  Also, the cpu usage hits 100% or close to and that is just for one user.  Any ideas on how to reduce the memory being used?  (When we created an extract the worksheets load very quickly although the extract took 30 minutes to be created.)  The data would need to be refreshed on a daily basis.