Re: Cube and Dimension Processing

Tech-Archive recommends: Fix windows errors by optimizing your registry

Hi Fernando,

Thanks for the response.

We have about 5 years of data in our datawareshouse.

Smallest fact table has 100,000 rows
Biggest fact table has 10 million rows.
Biggest dimension table has 500,000.

We are planning planning to populate data warehouse and process cubes
every hour. So the processing should not take more than few minutes.

One more question: if we do 'Process Full' on dimension, is it
necessary to do 'Process Full' on dependent cubes? Will 'Refresh' or
'Process Add" work in this situation?


Thanks
Mahesh



On Jun 28, 7:18 am, Fernando Ponte <ponte1...@xxxxxxxxx> wrote:
Hello Mahesh.
Answers below.

Remember. In all cases the dimension processing option has a direct
effect in the fact (cube) processing option.
I'm not considering the Proactive Cache option in the answers
(SSAS2k5).

What will be the best processing option under these four situations?

1.
Dimensions: addition of new members but no changes in existing members
Fact: new records but no changes in existing records.

Dimensions: Incremental
Fact: Incremental

2.
Dimension: addition of new members but no changes in existing members
Fact: new records as well as possibility of changes in existing
records.

Dimensions: Incremental
Fact: Refresh

3.
Dimension: addition of new members as well as changes in existing
members like moving under new parent.
Fact: new records but no changes in existing records.

Dimensions: Refresh
Fact: Refresh/Full

4.
Dimension: addition of new members as well as changes in existing
members like moving under new parent.
Fact: new records as well as possibility of changes in existing
records.

Dimensions: Full
Fact: Full

Will the processing option be different if the changes in existing
records are extremely rae vs very frequent changes? Any response will
be very much appreciated.

What is the volume of data to be processed?
How much time do you have to process it?

These questions will have more influence in your processing options.
Hope this help in some way.


.

Relevant Pages

  • Re: Cube and Dimension Processing
    ... In all cases the dimension processing option has a direct ... addition of new members but no changes in existing members ... Dimension: addition of new members but no changes in existing members ... members like moving under new parent. ...
    (microsoft.public.sqlserver.olap)
  • Re: Cube and Dimension Processing
    ... make sure your dimension use the flexible relationship between your members if you anticipate some changes between these attributes. ... remember if you have a rigid relationship between 2 attributes and if there is a change in the content, you must do a full process (of the dimension and also the cubes). ... In all cases the dimension processing option has a direct ...
    (microsoft.public.sqlserver.olap)
  • RE: Calculating Max Usage with OLAP
    ... to answer your question about how to only build aggregations at ... dimension and then looking at the Advanced tab in the properties pane). ... The combined number of members on both these ... > Max Used Licenses means –how many concurrent licenses have been used during ...
    (microsoft.public.sqlserver.olap)
  • Re: Problem with Member Counts and applying filters using MDX in M
    ... Now i am trying to count the "members" in this big dimension based on ... various possible filters that the user might apply. ... So far i am able to get the count of members on this big dimenion with one ... Here is the MDX i am trying.... ...
    (microsoft.public.sqlserver.olap)
  • Re: MDX Compatibility Property
    ... It looks like a bug. ... Please do not send email directly to this alias. ... The dimension browser forms an integral part of our ... have hidden members work ok! ...
    (microsoft.public.sqlserver.olap)