SQL Server Performance

Cube Partitioning

Discussion in 'Analysis Services/Data Warehousing' started by DBA_Ashfaq, Jan 4, 2007.

  1. DBA_Ashfaq New Member

    Hello Everyone

    Last year we started analysis services and i have data in my cubes. now the year 2006 end. i want to seperate data i mean creating partitioning. i am not sure what to do if i create partitioning then how i will address 2007 data in cube. do i have to recreate all cubes and dimensions for 2007 or is there any way that i can copy the structure of my 2006 cubes and dimensions and just change there name or whatever. please post if you have any idea.
    also you can e-mail me at dba.ashfaq@gmail.com
    i will appreciate your replies.

    thanks


    Ashfaq
  2. ranjitjain New Member

    Hi Ashfaq,
    Cube which you need to partition will be addressing to time dim as well right?
    So when you create partion you can specify the time slice value i.e. 2006/2007.
    Each cube has one partition by default and if you want to place your data in more than one partition it needs to be categorised. You can do this by specifying the Year value.
    After that if you want you can copy the already designed aggregations from old to new partition.

    So there will not be two different cubes but one cube with two partition will exist.
    You will continue to refer to this cube in MDX in same way.
    Cube internally will handle partitions differently.
    If you ask data for 2006, cube will read it from 2006 partion and if you analyze any data related to 2007 year, cube will read it from other partion and so on.

    In short, partitioning can not be expained in just one post, you need to read more articles.
    I will try and post any good links found, related to this topic
  3. DBA_Ashfaq New Member

    if you have please post it.
    thanks


    Ashfaq
  4. ranjitjain New Member

Share This Page