July 6, 2018 at 12:11 pm
Dear all,
I have speaking with people that defend that even using tabular model we could and should instead of using one start schema with some facts, use isntead multiple start schemas (according to the granularity of users needs).
Arguments is that it is better for performance and access management. As each group of users can use is start schema according to thir granularity and have access only to it.
According to what I saw , it seems that we can even have just one datamart with multiple start schemas Vs multiple start schemas each one representing a data mart.
Can you please point me to a document that has this explanation according to tabular? I though that this was only possible for multidimensional . But it seems tabular can and should also benefit from it.
I guess that this is the difference between imon and kimball?
Thank you.
July 7, 2018 at 2:26 am
river1 - Friday, July 6, 2018 12:11 PMDear all,I have speaking with people that defend that even using tabular model we could and should instead of using one start schema with some facts, use isntead multiple start schemas (according to the granularity of users needs).
Arguments is that it is better for performance and access management. As each group of users can use is start schema according to thir granularity and have access only to it.
According to what I saw , it seems that we can even have just one datamart with multiple start schemas Vs multiple start schemas each one representing a data mart.
Can you please point me to a document that has this explanation according to tabular? I though that this was only possible for multidimensional . But it seems tabular can and should also benefit from it.
I guess that this is the difference between imon and kimball?
Thank you.
As always, the answer is "it depends".
😎
How much data overlap would there be if separated into multiple star schemas?
Tabular models do not necessarily have to be star schema models and access can be controlled by row level access using i.e. roles.
July 7, 2018 at 10:03 am
I don't expet to havemuch more data. but more data will needed.
Solution is composed by:
1 staging area , one data mart one tabular cube.
We have one start schema with some facts.
We expect our DBs to go above 1 TB of information.
I believe that I can have (even wit tabular) multiple start schemas, each with one with one or two facts. and have just one data mart to old it all. or, have one data mart per each of the start schemas.
But I don't understand the advantages of this. In tabular should I have just one DM ?
Thanks
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply