none
Fact table design RRS feed

  • Question

  • I have an audit system db which contains audit data of various inquiries into the system.
    I want to create a DW architecture for it, I am a bit confused with the Fact table structure. Since my audit system has queries with character data how should i go about building the fact table [coz from what i understand fact tb only consists of numeric aggregates]

    What i am trying to achieve is a auditor can open excel and query my DW with different dimensions
    persons name/DOB inquiry dates etc....

    TIA
    SV
    Monday, April 6, 2009 2:26 PM

Answers

  • Hi You can define a named calculation in your dsv  to put one against each row and get data like the number of requests per month and also you can define the cout of presons who have requested. The dimensions can be see across this numeric data or if you are not interested in the numeric data you can just use to display the databased on this numeric column which will be dummy
    Hemanth Kumar
    Monday, April 6, 2009 6:01 PM

All replies

  • You're probably looking at a factless fact table where your only metric is the event count. Each row in the fact table counts as a single event. Then the analysis of your business domain proceeds like any other business process. What actors are involved? What attributes need to be captured? What are the relationships between those dimension attributes? etc.
    Monday, April 6, 2009 5:28 PM
  • dimensions:
    inquiry date
    Region


    ER is simple
    benefit table
    person table
    requesting application table

    each person has a benefit
    each requesting application seeks info on person or group of persons

    we want to create cubes so that we can drill based on
    1. Requesting App
    2. Benefits
    3. Time
    4. Demographics

    TIA



    SV
    Monday, April 6, 2009 5:36 PM
  • Hi You can define a named calculation in your dsv  to put one against each row and get data like the number of requests per month and also you can define the cout of presons who have requested. The dimensions can be see across this numeric data or if you are not interested in the numeric data you can just use to display the databased on this numeric column which will be dummy
    Hemanth Kumar
    Monday, April 6, 2009 6:01 PM