How to avoid snowflaking?
2 posters
Page 1 of 1
How to avoid snowflaking?
Hello
We have a FactDailyBalances snapshot table surrounded by DimAccount and DimDate dimensions. There are other dimensions too and we have avoided snowflaking i.e. we have a simple star schema.
The scenario I am grappling with is this: DimAccount has an OpenedDate attribute which at the moment is not linked to DimDate. Because of this, using the Cube, it is not possible to use a lot of the really useful date related attributes on DimDate. If I link OpenedDate to DimDate, I will end up with a snowflake:
DimAccount <---- FactDailyBalances ---> DimDate
|-----------------------------------------A
Is there a way achieving what I need without snowflaking?
We have a FactDailyBalances snapshot table surrounded by DimAccount and DimDate dimensions. There are other dimensions too and we have avoided snowflaking i.e. we have a simple star schema.
The scenario I am grappling with is this: DimAccount has an OpenedDate attribute which at the moment is not linked to DimDate. Because of this, using the Cube, it is not possible to use a lot of the really useful date related attributes on DimDate. If I link OpenedDate to DimDate, I will end up with a snowflake:
DimAccount <---- FactDailyBalances ---> DimDate
|-----------------------------------------A
Is there a way achieving what I need without snowflaking?
amir2- Posts : 29
Join date : 2010-07-29
Re: How to avoid snowflaking?
I am not a proponent of snowflake, which I think is obvious from prior postings. However, if you need the attributes, this would be a case where a snowflake makes sense. I say this because something like open date is static and it doesn't make much sense to include it as an FK on all facts that also reference account. The other alternative would be to place selected date attributes in the account dimension, but there is always the risk you miss something the user wants to use, and it kind of looks bad...
Similar topics
» Avoid role play dimension in DW
» Question on joining dimensions to avoid possible snowflake
» Using larger data type to avoid sum overflows?
» Time Dim Snowflaking
» SNOWFLAKING
» Question on joining dimensions to avoid possible snowflake
» Using larger data type to avoid sum overflows?
» Time Dim Snowflaking
» SNOWFLAKING
Page 1 of 1
Permissions in this forum:
You cannot reply to topics in this forum