Skip to Content
author's profile photo Former Member
0
Former Member

Basic modeling question

Posted on Nov 27, 2005 at 07:32 PM 10 Views

Hi all,

I would like to allow reporting of some system logs stored in database tables (header and item) in BW.

I have the choice of storing it in an ODS object or an InfoCube.

Storing them in an ODS object is the first option I could think of, but since I expect users to slice and dice with the data in reports, I am tempted to use an InfoCube with appropriate dimesions designed.

Having said that, I am not still not sure how slice and dice with an InfoCube is faster in this scenario as I will be able to store all data in one flat table in an ODS. With some secondary indexes I should be able to get good performance.

I am not completely sure which approach is correct. Could some one throw some more light on this?

Regards,

Girish

Add a comment
10|10000 characters needed characters exceeded

Assigned Tags

Related questions

1 Answer

  • author's profile photo Former Member
    Former Member
    Posted on Nov 27, 2005 at 11:40 PM
    0

    HI girish,

    Reporting on ODS should be avoided as far as possible for perfoarmance reasons. For the time being you can use both a ODS and a cube. Use ODS to update the cube by mapping one is to one. Try reporting on both, probably you can do some R & D ofurther and get the required design.

    Regards

    Sriram

    Add a comment
    10|10000 characters needed characters exceeded

    • Former Member
      Nov 28, 2005 at 01:50 AM

      The decision to do reporting out of an ODS or cube really depends on a ouple of things. I don't think that it needs to be avoided, and could be your best option, but really depends on the data and the queries.

      If the data will be looked at in summarized form only, cube would be good - that's one of the cube's strengths. If you need to always see the detail, an ODS might be best. Another factor that might lean me to choose an ODS is if the querying is very specific, e.g. just one or only a few queries, and the selection criteria (so you can build indices)is fairly constant.

      A cube might be better if you have many characteristics and the selection criteria will vary a lot - having bitmap indices and using a star transformation (I'm assuming Oracle here) can really help when the queries tend to use a variety of criteria from one run to another.

      Nothing says you can't use both, if you need to suport both summarized and detail querying, and use RRI to jump from a summarized result from the cube to underlying detail in the ODS.

Before answering

You should only submit an answer when you are proposing a solution to the poster's problem. If you want the poster to clarify the question or provide more information, please leave a comment instead, requesting additional details. When answering, please include specifics, such as step-by-step instructions, context for the solution, and links to useful resources. Also, please make sure that you answer complies with our Rules of Engagement.
You must be Logged in to submit an answer.

Up to 10 attachments (including images) can be used with a maximum of 1.0 MB each and 10.5 MB total.