Forum Discussion
Power BI vs Data Warehouse
Power BI should not be used as a data warehouse. It works well with lots of data sources, including a data warehouse, but lacks some key functionallity around change tracking and incremental updates. These are complementary technologies.
- Anonymous9 years agoNot applicable
Thanks Phil.
I agree, Power BI worked great with data warehouse, except perhaps for not being able to combine data from the DWH which you might quickly want to mash up with other data and drill through - we still cant drill through to row detail [see records menu] held in SSAS Tabular from Power BI (which you can do if you import data into Power BI).
Additionally with direct query we can query underlying data in SQL Server and build a model in Power BI rather than needing to build the model in SSAS.
At the moment it feels like Power BI development is outpacing SSAS Tabular which is making some SSAS Tabuar functionality redundant.
Do we have a roadmap about how SSAS Tabular will be developed and integrated with Power BI, or will Power BI eventually replace SSAS Tabular?
Thanks!
Pbix
- MattAllington9 years ago
Community Champion
As Phil has said, a Data Warehouse and Power BI are complementary. A Data Warehouse is a storage database, Power BI is a reporting database.
Anonymous wrote:I agree, Power BI worked great with data warehouse, except perhaps for not being able to combine data from the DWH which you might quickly want to mash up with other data and drill through
I don't understand the point you are making here. Power BI is explicitly designed to get data from anywhere and mash it up - that's what it does.
- we still cant drill through to row detail [see records menu] held in SSAS Tabular from Power BI (which you can do if you import data into Power BI).
Yes you can. SSAS Tabular has an identical reporting engine underneath as Power BI. If you write the identical database in Power BI and SSAS Tabular you will get identical results. If you can't drill down to see the row level detail then it is related to your design, not SSAS Tabular
Additionally with direct query we can query underlying data in SQL Server and build a model in Power BI rather than needing to build the model in SSAS.
Direct Query does not have a data model in Power BI. You either connect to SQL Server, or SSAS and use the source to determine what you can see, or you build a model in Power BI - you can't do both at once.
At the moment it feels like Power BI development is outpacing SSAS Tabular which is making some SSAS Tabuar functionality redundant.
As I mentiond above, SSAS Tabular and Power BI use the same underlying reporting engine (Vertipaq/Power Pivot). There is very little development in this technology in either product. The development of Power BI is mainly in the area of data import (Power Query) and visualisations.
What is becoming redundant in SSAS Tabular?
Do we have a roadmap about how SSAS Tabular will be developed and integrated with Power BI, or will Power BI eventually replace SSAS Tabular?
SSAS Tabular is already integrated with Power BI. No Power BI will not replace SSAS Tabular. SSAS Tabular is only 1/3 of what Power BI is. Power BI has a reporting database (Power Pivot), an ETL tool (Power Query) and a visualisation/reporting tool (Power BI Visuals). SSAS Tabular is just the first one (effectively Power Pivot for Enterprise).
- Anonymous9 years agoNot applicable
Hi Matt,
Great, thanks for your thoughtful explanation of SSAS vs Power BI - just picking up on your points:
MattAllington wrote:I don't understand the point you are making here. Power BI is explicitly designed to get data from anywhere and mash it up - that's what it does.
V much agree - have I missed something though? Once I've connected to SSAS Tabular I can't see how to use Desktop Designer to mash this up with other sources in reports/datasets? We can mash SSAS Tabular data up in the Power BI service though. Great if I'm wrong on this though! :)
Yes you can. SSAS Tabular has an identical reporting engine underneath as Power BI. If you write the identical database in Power BI and SSAS Tabular you will get identical results. If you can't drill down to see the row level detail then it is related to your design, not SSAS Tabular
Again, have I missed something basic here? I recognise that Power BI uses the same engine but when I import non-SSAS Tabular data into Power BI and visualise this in a chart I can right click on the chart, select 'see records' and I can then see underlying row-level data. If I direct query from SSAS Tabular, put Tabular data in a chart then and then right click on it I can't see the 'see records' option any more. Again great if I'm missing something obvious!
As I mentiond above, SSAS Tabular and Power BI use the same underlying reporting engine (Vertipaq/Power Pivot). There is very little development in this technology in either product. The development of Power BI is mainly in the area of data import (Power Query) and visualisations.
What is becoming redundant in SSAS Tabular?
Yes, redundant was probably the wrong choice of words as it uses the same engine as Power BI. I suppose I meant that Power BI seems to offer lower barriers of entry to most users than compared with SSAS Tabular. From my novice perspective, it feels like Power BI has a USP whereas SSAS Tabular has a bit of an identity crisis between its older OLAP brother and younger Power BI brother.
No Power BI will not replace SSAS Tabular. SSAS Tabular is only 1/3 of what Power BI is. Power BI has a reporting database (Power Pivot), an ETL tool (Power Query) and a visualisation/reporting tool (Power BI Visuals). SSAS Tabular is just the first one (effectively Power Pivot for Enterprise).
Great, thanks for summarising this so concisely.
Thanks for your perspective - really appreciate it. :)
Pbix
- Phil_Seamark9 years ago
Microsoft Employee
I'm not an MS insider so can't speak for sure. But remember that Power BI is SSAS tabular under the covers (all be it a cut down version running in diskless mode). The main difference is that PBI desktop is on a more regular release schedule. So I would expect to see these feature appear in future releases of SSAS tabular. Perhaps SSAS Azure will see them first.