Qlik Community

QlikView Documents

Documents for QlikView related information.

Slowly Changing Dimensions

MVP
MVP

Slowly Changing Dimensions

Hi All,

The content in the below URL helps in understanding the Slow Changing Dimensions concepts of Datawareshousing.  Hope this helps.

Source:What are Slowly Changing Dimensions?

Regards,

Jagan.

Comments
jvitantonio
Valued Contributor III

Hi Jagan,

I find it a bit pointless to copy something from the internet and paste it into a Word document to then share it in the community.

Although the information provided is interesting, you are not adding any value to the source. Providing the source URL or the Word document in this case is the same.

It would be a lot more interesting to the community (or at least to me), if you can create content to this same document by explaining how you would implement  the different techniques in Qlikview / Qlik Sense.

Thanks,

JV

MVP
MVP

Hi Juan,

Thanks for your feedback.  I gone through this content and thought it is worth sharing in community, especially it is very useful for beginners in Qlikview/Qliksense.  May be if I get some time I will update this document with the implementation in Qlikview.

Regards,

Jagan.

maxim_senin
Contributor III

Hi guys,

I'm sure that here first thing which must be understood - whether it's necessary to support a dimension (which changes over time) as slowly changing dimension (SCD) or it's possible to ignore fact of changing.

As for my experience (retail, logistics and wholesales) about 25-50% of dimensions can be SCD, but only 5-10% of them are supported as SCD. Sometimes (or even frequently) it's just not needed at all.

It's called "Type 1" in the article and I believe that should mention an estimation of a necessity of SCD in general.

Best regards,

Maxim

hugmarcel
Contributor III

Hi Maxim

I agree to you regarding use of "Type 1" rule. During my career as a BI-specialist, I have been working with different methods.

It turned out that using higher SCD types can become very complex for end users and business employees, as they usually do not understand the impact of historical dimension data to their reports.

E.g. think about security: Having security rules depending on hierarchical organisations: Shall organisation structures be kept historical or not? And what's the impact on data visibility if a business member changes organisation, department etc.?

If you want to implement higher levels properly, a precise definition regarding SCD has to be made on EVERY dimension needed. And it must be carefully explained with proper EXAMPLES, etc.

My experience is that you will end up with lots of (desperate) end user questions if you are making heavy use of higher type SCD-methods.

Marcel

Version history
Revision #:
1 of 1
Last update:
‎04-29-2015 03:58 AM
Updated by: