Power BI - Cube like data access - Performance Issue - Snowflake on cloud as backend data store

Viewed 31

I have a cube like use-case which I want to handle in Power BI. My FACT table is having huge data for every business day. My model is a perfect STAR schema model. But I want first 5 days' data to be available in Power BI as import mode (for better performance) and rest of the data to be available as direct query. I am using snowflake on cloud for storing back-end data. I already tried direct query mode for FACT data as a whole and performance is not that impressive. Please suggest if there is an better option to handle this in Power BI.

1 Answers
Please share what are the dimensions you are trying to display.
as I see, you are trying to display some value ( earning or any continuous 
variable) over last 5 days , along with other dimensions involved ( as its in 
OLAP or any MDDB).

You can take 2 approaches. 

you can introduce a variable to only  pick last rolling 5 days of data at the x 
axis

Approach1 : 
in a PBI column chart 
x axis = Dates 
y axis = the value that you want to display 
add slicers to reflect other dimensions ( such as geography , designation etc 
other dimension. 

Approach2: ( good for 3 dimensions , more dimensions possible but makes display 
messy. 
Again pull a PBI column chart. 
x axis = Date 
y axis  = the value that you want to display  
small multiples = combination of other dimensons
Related