Forum Discussion
Import or Composite?
I need historical data for the calendar year and the last three calendar years.
Assuming the 2GB data comprise data for the 4 calendar years, and that PowerBI premium is used, Import mode should be fine, right?
What I wanted to understand further is, if Import mode is used, will that affect the live alert for the late order (Although I don't think it matters since Import mode allows daily refresh, and the late order is dictated by whether Shipped Date > Required Date, which is a daily incremental figure)
Is the requirement of the report data being updated as of 7am daily, only able to be fulfilled via Import mode and no other modes?
I need some thorough explanation here to tackle a tricky question like this.
YoungLearning Import should be fine, especially if you are mindful of which columns you are importing from the source to ensure that your semantic model is as small as possible per the requirements. Also, the requirement for fast response times would also skew towards import.
Hi YoungLearning ,
First of all, thank you YoungLearning and Greg_Deckler for the quick reply. I have some thoughts to add:
Using a role-playing dimension, as described in option A, can indeed be a valid approach for handling multiple date columns like OrderDate, ShipDate, and RequiredDate. This method allows you to create multiple relationships between the date table and the fact table, with one active relationship and others inactive. You can then use DAX functions like `USERELATIONSHIP` to activate the necessary relationship in your calculations.
However, the correct answer given is option C, which involves using the Auto date/time option in Power BI Desktop. This approach simplifies the model by automatically creating date tables for each date column, which can be useful for quick and straightforward reporting.
In summary, while option A is a valid method and can be used effectively, option C is chosen here likely because it directly addresses the requirement to simplify the model and make it easier to manage date-related calculations without manually creating and managing multiple relationships.
Best Regards,
Neeko Tang
If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.
4 Replies
This is the full specs
YoungLearning Import should be fine, especially if you are mindful of which columns you are importing from the source to ensure that your semantic model is as small as possible per the requirements. Also, the requirement for fast response times would also skew towards import.
Follow up question:
The fact table has order date, shipping date and required date... Would using role playing dimension, as in option A suffice? Late shipping requires shipping and required date information.
Hi YoungLearning ,
First of all, thank you YoungLearning and Greg_Deckler for the quick reply. I have some thoughts to add:
Using a role-playing dimension, as described in option A, can indeed be a valid approach for handling multiple date columns like OrderDate, ShipDate, and RequiredDate. This method allows you to create multiple relationships between the date table and the fact table, with one active relationship and others inactive. You can then use DAX functions like `USERELATIONSHIP` to activate the necessary relationship in your calculations.
However, the correct answer given is option C, which involves using the Auto date/time option in Power BI Desktop. This approach simplifies the model by automatically creating date tables for each date column, which can be useful for quick and straightforward reporting.
In summary, while option A is a valid method and can be used effectively, option C is chosen here likely because it directly addresses the requirement to simplify the model and make it easier to manage date-related calculations without manually creating and managing multiple relationships.
Best Regards,
Neeko Tang
If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.
