8 Replies Latest reply on Sep 3, 2013 2:15 AM by kettan

    Server Data: When is it better to use packaged files than server extracts?


      I expect packaged workbooks to (1) perform better and be (2) less time-consuming to manage in our setting than server extracts. An added benefit is that they are (3) ready to be downloaded for offline use by dashboard consumers.

      Therefore I seriously consider using packaged workbooks for most of our dashboards and limit server extracts to a few large* data sets until a better solution is available, such as Dimensionally Modeled Extracts, and also easy access to Report Metadata.

      * large is 10-20M rows


      I am not happy with my choice, because it is not a good choice, just the least painful I can think of. Therefore I welcome any advise about minimizing the pain of managing data used in Tableau Server. This could include live connections to our data warehouse run by SQL Server 2012 Enterprise.


      CriteriaTWBXTWB + Extract

      Good; includes only columns needed

      Intermediate; has more columns than needed in each singular TWB


      Easy to re-use by modifying underlying queries. Difficult, because number of files is high; Difficult, because we cannot cascade changes & corrections of underlying queries

      Difficult to re-use, because (1) it is difficult to know which other TWBs use the same extracts, (2) and a pain to change underlying queries


      Bad, because there is no re-use of data. It also costs much more time to transfer data to Tableau Server.

      Good, because sharing extracts minimizes disk space usage & extract refresh time


      Easy, because refreshed workbooks can be downloaded from the server ready for offline use

      Difficult to save as packaged workbooks


      I rated the LIVE, EXTRACT, and PACKAGED data options by the most important criteria and came (not surprisingly) to the conclusion that using packaged workbooks will be the least painful route to take.

      Live vs Extract vs Packaged.png