none
SSAS Process Default option for dimension

    Pregunta

  • Hi,

    I am trying to process the SSAS dimensions with option "Process Default" which should ideally choose the best processing option to bring the dimension to a fully processed state.  But when I use that option it doesn't even bring in the new members added to my dimension.  Is it a bug? or am I missing something here?

    Any help is much appreciated.

    Thanks,

    Vinaya

    jueves, 21 de junio de 2012 23:09

Respuestas

  • Hi Vinaya,

    Process Update would be the way to go. Process Default will not add new members if the dimension is already in a processed state. The general process I usually employ is the following:

    • Perform a Process Update on dimensions. - Aggregations associated with flexible relationships will be dropped so your users may experience performance degradation, especially on a large cube. The alternative is to perform a Process Update and check the box to also process affected objects. This option essentially processes your flexible aggregations along with the member changes in your dimension. However, with a very large cube doing so may increase your dimension processing (and memory consumption) significantly.
    • Perform a Process Data operation on measure groups which had new data added.
    • Perform a Process Indexes on all measure groups. - This step rebuilds all aggregations associated with flexible aggregations and all aggregations on previously processed measure groups and all aggregations on rebuilt measure groups.

    HTH, Martin


    http://martinmason.wordpress.com

    • Marcado como respuesta Vinaya Ravi martes, 26 de junio de 2012 23:02
    martes, 26 de junio de 2012 12:54

Todas las respuestas

  • "Process Default" will bring it to a fully processed state but that doesn't mean that it will be in a consistent state with your relational source table. If your dimension is already in a processed state, it won't do anything. "Process Update" is the process that you're looking for which will bring in new dimension members and update attribute relationships, failing in cases where a rigid relationship incurs changes. http://msdn.microsoft.com/en-US/library/ms174774(v=sql.105).aspx

    HTH, Martin


    http://martinmason.wordpress.com

    viernes, 22 de junio de 2012 2:58
  • Thanks Martin.  But I would like to have an option where in, I need to do minimal processing to bring the dimension to processed (whatever change it may be, a structural change or new members added/removed etc.).

    I thought "process default" does that, if its not will "Process Update" work perfect for this situation?  I dont want to do a full process of dimension everytime.  So can you please suggest a better way to process Dimensions and related measure groups that will do only necessary processing to get the data up to date?

    Thanks,

    Vinaya

    lunes, 25 de junio de 2012 23:22
  • Hi Vinaya,

    Process Update would be the way to go. Process Default will not add new members if the dimension is already in a processed state. The general process I usually employ is the following:

    • Perform a Process Update on dimensions. - Aggregations associated with flexible relationships will be dropped so your users may experience performance degradation, especially on a large cube. The alternative is to perform a Process Update and check the box to also process affected objects. This option essentially processes your flexible aggregations along with the member changes in your dimension. However, with a very large cube doing so may increase your dimension processing (and memory consumption) significantly.
    • Perform a Process Data operation on measure groups which had new data added.
    • Perform a Process Indexes on all measure groups. - This step rebuilds all aggregations associated with flexible aggregations and all aggregations on previously processed measure groups and all aggregations on rebuilt measure groups.

    HTH, Martin


    http://martinmason.wordpress.com

    • Marcado como respuesta Vinaya Ravi martes, 26 de junio de 2012 23:02
    martes, 26 de junio de 2012 12:54
  • Thanks a lot for your help :)
    martes, 26 de junio de 2012 23:02