Power BI User on 09 Jul 2015 12:35:38
Azure SQL and Azure DW Direct Query datasources are great, and this is most probably the future. There is however a requirement for a real-time semantic model in the middle. For example, this will be great for: captions/translations, attribute order by, hierarchies, images, relationships defined outside of the db, ... And most importantly that could provide a multidimensional query language (MDX and/or DAX). It could be Tabular in Azure or it could an in-database semantic model like SAP HANA?
- Comments (20)
RE: Direct Query Semantic model
When dealing with large data sets this is a huge requirement. Awesome!!!
RE: Direct Query Semantic model
This limitation also impact me. A simple addition is not possible.
Moreover I think it is something that can easily be added to a futur Power BI version.
RE: Direct Query Semantic model
Direct Query support with the ability to blend this source (summarized by source system) is crucial for its adoption as an Enterprise - Self Service Reporting and Dash boarding tool. Microsoft Power BI Team, request you to consider this as soon as possible.
RE: Direct Query Semantic model
Although this hasn't been a huge ask from most of my customers, I can definitely see this coming up more in the future. As many of our customers adopt the self-service BI model for their organizations, I can visualize this becoming more and more of an ask in the future as the volume of data continues to increase.
RE: Direct Query Semantic model
Agree with everyone here - we must combine multiple connections to get live and historical data into PBI.
RE: Direct Query Semantic model
I need LOOKUPVALUE to work with Direct Query mode to implement access rights
RE: Direct Query Semantic model
It would be great to use excel in direct query mode!!!!!!!!!!!!!!
RE: Direct Query Semantic model
With 20 years experience working with tools having semantic layers like Cognos and Business Objects, I would say that semantic layers are often more trouble than they are worth (except for well designed OLAP models). For example, whilst it may seem great to be able to join data from completely separate databases it can lead to serious performance issues as usually full tables from both databases have to be downloaded to the (usually underpowered) report server. This is all hidden from the user who has no idea why the reports run so slowly. Therefore I would hope for something smarter. Something like the ability for power users to build multidimensional models which are then pushed down to SSAS Tabular which then hopefully could be partially populated on-demand (instead of requiring a full refresh) essentially caching it for subsequent users. Loads for frequently used Tabular models could then be streamlined, consolidated and formalised by IT on a continuous improvement basis.
RE: Direct Query Semantic model
Cannot upvote this enough. This is a huge gap right now that needs to have first priority from a dev perspective. There is no way to have governance over measure calculations for measures that have very well defined change control. If I'm using PowerBI to do financial reporting, I need to be able to control how measures are calculated, and the best way to do this is to have a published semantic model hosted either in Analysis Services or in something like SAP HANA.
RE: Direct Query Semantic model
This is really THE thing that our business needs! I'm working on the implementation of Power BI in our company and it's just not convenient to have an intermediate SQL database.