SQL Server Performance

Olap Process

Discussion in 'Analysis Services/Data Warehousing' started by sramesh, Apr 25, 2006.

  1. sramesh New Member

    Hi all
    ( i am new to olap)
    this is my query
    select unitname, qty, productioncost, productionamount, date from production
    Tho 5000 56 53444 01/04/2006
    Tho 5000 55 53444 02/04/2006
    Tho 5000 54 53444 03/04/2006
    kat 1000 56 53444 03/04/2006
    .......
    Tho 5000 56 53444 25/04/2006
    kat 2000 50 51254 25/04/2006
    i have processed this query.
    now my cube have upto date record.
    one user updated a record for date 10/04/2006.
    i want reprocess (incremental mode) and datas fetch from 10/04/2006.
    what happend now?
    existing records (10/04/2006) onwards overright or else?
    if doubled records, i want delete 10/04/2006 onwards record before reprocess.
    and also i need dso object reference or sample applications for vb.



    Thanks
    S. Ramesh
  2. ranjitjain New Member

    Please clearly write your problem, it's really confusing.

    The data which you do not want in your cube then you just need to delete them from source table i.e. your dim table/fact tables.

    You need to use DSO(Decision support object) to refer to any MSAS OLAP activities.

    If your data has been changed then you just need to reprocess the cubes so that new data will get loaded, full process in case if any previous record is been modified.
  3. RGKN New Member

    I thinks that Ramesh is saying that

    They have a fact table which has had a record updated (dated 10/04/2006) He then wants the data updated incrementally only changing data after that date. He has run incremental load and has found that all the data has duplicated.

    Do correct me if I misunderstand, Ramesh.

    I think the difficulty you are having, Ramesh, is that you have misunderstood what the incremental load does. The incremental load will only add data to your cube, it does not update data. In order to update the data you need to run a full process. The incremental load would only be useful if you are wanting to add new data to the cubes. For example in a financial system where you want to add new transactions you would have only the new transactions in the fact table you load from, then run an incremental load and those will be added to the cube. If you run the incremental load from a table that still holds the same data as before, the cube will have duplicate data added to it.

    Regards,

    Robert.



  4. sramesh New Member

    Thanks Robert.
    now i cleared. is it possible to remove some set of datas from the cube?
    i mean i want remove datas after 10/04/2006. is it possible?
    for a example because full process take 30 minutes. so i want avoid process time.




    Thanks
    S. Ramesh
  5. sramesh New Member

    Hi
    one more point,
    for a example,

    insert tempcube
    select * from cubeproduction where pdate< '10/04/2006'

    insert tempcube
    select * from production where pdate >= '10/04/2006'

    cubeproduction = tempcube
    IS IT POSSIBLE IN CUBE?
    I AM NEWER IN CUBE. I NEED THIS TYPE OF MDX QUERY ANY BODY GIVE EXAMPLE OR HINT.



    Thanks
    S. Ramesh

Share This Page