What is incremental aggregation/2. What is rapidly changing dimension?3. How can we merge the two dimensions?4. Can we edit the lov’s? How many ways we can edit?5.What is the difference between rank and dense rank?6. Tell me the process of creating universe?7. How can you send webi reports to the repository?8. Have u created folders?9.What is the object? Difference between XI R2 and 6.5?10. How can we migrate from development server to productions server?

Showing Answers 1 - 2 of 2 Answers

Ill answer the first three. An incremental aggregation (I think because I have not exactly used the term) is an aggregate in which the new aggregate rows are appended (added to the back of) previous aggregate rows, or earlier aggregates. It is in contrast to a full-refresh where the whole aggregate is replenished (i.e., deleted and rebuilt).

A rapidly changing dimension has two possible descriptions: a dimensions that changes a lot (almost as frequently as transactions happen); or a dimension that grows to a humongous size. Any changing (or historical) dimension where the changes are kept as changes is inherently incremental -- the changes are added to the end of the existing dimension records. Kimball differentiates 3 types of slowly changing dimensions. Type 1 overwrites changes. Type 2 maintains a history. New rows are added. That is what I just described above. Type 3 maintains (in a single row) the current value of something and the last value of it. Type 2 dimensions are inherently incremental.

How do you merge incremental aggregation and rapidly changing dimensions? Several ways I guess: one is instead of keeping all this rapidly changing data in all its details, that you summarize it. This summary or aggregate rows are appended to the end of the previous aggregate rows. Often after such data is summarized, the older detailed (rapidly changing) data is rolled off the database (deleted, but the summaries are kept). Why is it rolled off? Because rapid changes can easily become many, many changes, which can become humongous, so they are summarized, the summaries are kept, the detailed are deleted or stored on something else, like tape. To reduce the amount of rapid changes, some fields can be redefined to value bands. E.g., instead of keeping someone exact income, you use ranges of values to replace them, such as income up to $50K, $50-100K, $100-150K, etc.

  Was this answer useful?  Yes


  • May 20th, 2012

Incremental aggregation concept is required in incremental loading. For example, lets say we receive sales data weekly and the same data is stored in some fact table with monthly granularity. We might want to load that monthly fact on weekly basis. So each time we will have to perform incremental aggregation thereby modifying the measures as until that week

  Was this answer useful?  Yes

Give your answer:

If you think the above answer is not correct, Please select a reason and add your answer below.


Related Answered Questions


Related Open Questions