[GitHub] [superset] wernerdaehn commented on issue #12729: [SIP] Support star schemas

[GitBox <gi...@apache.org>]

wernerdaehn commented on issue #12729:
URL: https://github.com/apache/superset/issues/12729#issuecomment-766685154


   @zhaoyongjie From a technical side, if all databases would support views with a virtual unlimited number of columns and do a perfect job from the SQL optimizer point of view, then yes. But the opposite is the case. Only Oracle And SAP Hana (Calculation Views) are even close to that and far from perfect.
   From a logical point of view, Superset wants to sum up measures and group by attributes. That is a dimensional concept. 
   Further more, if the tool knows what a fact table is and what a dimension, we have much more options later. For example you might have an ORDER and DELIVERY table, both have a shared dimension CUSTOMER. In a single-table model the customer table with its 500 columns must be defined multiple times whereas with a shared dimension just once. And in the dashboard, when you want to view the order amount, shipped amount and order backlog, a shared dimension allows to filter all measures at once. You know what is shared and how.
   Also for the users the selection of columns is much easier when they are grouped. The customer master dimension obviously provides all columns that are customer related, the material all material related fields. In a single table model you would have 300 columns from the one and 300 columns from the other. And if customer can be filtered on sold-to, ship-to and bill-to basis, you would have the 300 customer fields three times.
   And finally, while today the semantic layer of Superset is very thin, it will grow. And then specifying the same information multiple times get even more of a burden.
   
   So no, this has nothing to do with OLAP, it is related to being a business intelligence tool.
   
   I see that as something fairly straight forward to implement and with lots of potential initially and later.


----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: notifications-unsubscribe@superset.apache.org
For additional commands, e-mail: notifications-help@superset.apache.org