Hello all,
This is a best practice question. Is it better to have dimensions coming into transformer as single data sources i.e 3 dimensions - 3 datasources plus the fact table, which then combine during build time. Or is it better to have a single data source coming from a star schema, with the relationships between the dimensions and fact table modeled in Framework manager.
I thought the second way works better in terms of cube build time. But I have been told that best practice is to bring each dimension in individually for a quicker read and build time???
So what is the best way for cube build time and best practice?
----------------------------------------
Another throw closer to a win ... ???
----------------------------------------
This is a best practice question. Is it better to have dimensions coming into transformer as single data sources i.e 3 dimensions - 3 datasources plus the fact table, which then combine during build time. Or is it better to have a single data source coming from a star schema, with the relationships between the dimensions and fact table modeled in Framework manager.
I thought the second way works better in terms of cube build time. But I have been told that best practice is to bring each dimension in individually for a quicker read and build time???
So what is the best way for cube build time and best practice?
----------------------------------------
Another throw closer to a win ... ???
----------------------------------------