0

We have 4 to 5 denormalized tables generated from SAP. How can we create a tabular data warehouse with denormalized tables? What is the recommended warehousing technology? All tables are pushed into SQL by SSIS by processing flat files from SAP RFC reports.

Sandra Rossi
  • 11,934
  • 5
  • 22
  • 48
Deva
  • 1
  • What type of users do you have? What client tools? How big are the flat files? How much RAM do you have on your servers to house this data? – GregGalloway Sep 11 '16 at 20:57
  • All users are BU heads and CXO's. We are using SSIS -> SSAS -> Power BI for visualization. We have 16 GB of RAM in all servers (Staging and Prod). Each text file is about max of 20 MB (approx 20K records with 30 fields on average). Thanks in Advance Greg. – Deva Sep 14 '16 at 16:20
  • 1
    sounds like you are on the right track. Tabular would be a good fit here. If reports never need to join tables then simply importing the tables into SSAS Tabular will work fine. If you need to join tables and cross-filter then you will want to consider building a Kimball style star schema with fact tables and dimensions. But is is great you are already loading the files to SQL before putting Tabular on top of SQL tables. – GregGalloway Sep 14 '16 at 23:51
  • Thanks Greg! But the other question is, Does Tabular model (SSAS) is really needed here? why Shouldn't we directly query the SQL tables? Here we don't have any facts and dimensions, so importing the data in tabular model and deploying it as a Cube/Model makes sense? My doubt is, we don't do much warehousing things here other than ETL and deploying tabular model, is it the right approach? – Deva Sep 16 '16 at 12:55
  • if you want drag and drop self-service reporting and want fast performance and want to bake in calculations then the cube is a good idea – GregGalloway Sep 16 '16 at 13:17

1 Answers1

0

It doesn't sound like you are using Tabular for any pre-aggregation, but rather a transport mechanism to get it to Power BI. You can model these tables directly in Power BI and get all of the benefits that you have to date. You would get additional benefits as well. Power BI would then allow the user to create their own measures, enhance the model with other tables, etc. If the issue is that you don't want people accessing this SQL instance directly, you have a lot of options there as well.

Michael buller
  • 566
  • 4
  • 11