Skip to main content
cancel
Showing results for 
Search instead for 
Did you mean: 

Earn the coveted Fabric Analytics Engineer certification. 100% off your exam for a limited time only!

jirineoral

Power BI – Live connection vs. Import, comparison and limitations

Some of the data sources have option both to import data as well as use live connection. But there is no easy method to change it once you already built your report. So you should rather carefully decide which method you’re going to use upfront. To make this decision right, it is good to know limitations of each method. Purpose of this blog post is to cover these limitations.

First what kind of license are you going to need? If you’ll use live connection, you’ll need Pro license for all end users consuming report/dashboard. If you’ll import data from on premises data source through gateway, you’ll need Pro License again. If your data are accessible without gateway, but you want to refresh them more often, then once a day, Pro License is required too. If you want to stay with free Power BI license, you’ll need to use data Import method only and avoid on premises data sources, refresh data maximally once a day.

 

Data import

Data import can be used against any data source type. Current Power BI service limitation published file size is 1 GB. Data are loaded to file using column store compression. If you’ll read for report purposes just data required for usage in visuals this is not limiting too much. Even with Power BI Pro license you are limited to 8 refreshes per day. Times can be scheduled either at full hour or half past full. Scheduled time is approximate. When will refresh actually happen depends on service utilization. If you need to refresh data, more often consider using live connection. When using import, data are stored in Power BI file/service. Therefore, there is no need to setup permissions on data source side (service account for load is enough) and you can share data publically or with people outside organization. On the other hand, all data are stored on Power BI. So if you have requirement for Row Level Security, you need to implement it on Power BI side. If you need same logic applied for multiple reports, you have to duplicate it in all the reports. Same applies to calculations, if you need same measure in multiple reports, you need to recreate it every time.

 

Live connection

There are more limitations for live connection in place. It doesn’t work against all data sources. Current list can be seen here https://powerbi.microsoft.com/en-us/documentation/powerbi-refresh-data/#live-connections-and-directq...

You are also limited to just one data source/database you selected. You can’t combine data from multiple data sources anymore. If you are connected to SQL Database, you can still create logical relationships between objects from that database as well as measures and calculated columns. When you are connected to SQL Server Analysis Services, you are limited just to report layout and even can’t make calculated columns or measures. When connected to SSAS, you could also reach some licensing limitations. Power BI issues DAX queries against SSAS and this isn’t supported in Standard edition of SQL Server before SQL Server 2016. When using live connection, users have to have access to underlying data source. This means you can’t share outside of your organization or publically. Other way around, when you set up security once, all reports using live connection to the same source have the same security model applied automatically. Similarly, you can prepare your data model carefully with all measures and calculated columns on data source side. All reports from the same data source can benefit of it. Refresh frequency is unlimited. Power BI simply shows data as they are in database. However, be careful using live connection against highly transactional OLTP systems. If you suffer from performance problems now, analytical workloads from Power BI won’t help it. Consider enabling features on data source side like Operational Analytics in SQL Server 2016.

 

Summary

Both methods have limitations, developer should be aware upfront. For import it is mainly refresh frequency, data size and necessity to duplicate same logic (calculations, security) in multiple reports. That could be helped by templates, but if you need to change logic of calculation, you need to change it in multiple reports again. For live connection, not all data sources are supported. If you pick one, then you’re limited just to that one data source for report. If logic is setup on source side, all reports can benefit. My personal preference is use live connection against SSAS models, have the model centralized and maintain logic in one place. For SQL Databases where I don’t have model, I prefer import and use Power BI as a model and vizualization.

Comments

Isn't importing from SSAS Tabular in Power BI similar to how SSRS report connected to SSAS models would work? SSRS honors the row-level security setup in the SSAS model so I assume when Power BI imports the data, it would also honor the row level security if authetication is set to impersonate the logged in user. Is that correct?

@vickyd - if you mean live connection to SSAS Tabular, yes it would read it from SSAS data model. It works against Multidimensional as well. But it doesn't work for import method. Import happens under one account you set up in data source, data are read to Power BI service during refresh and stay there. End user isn't authenticated by back end database when using import method Jiri

Hello Jiri, I have one problem with refreshing live Azure AS conection, could you help me please?
It means that in Visual Studio (SSDT) I see in model after processing table 42 rows trasnferred (I deleted one row in DTB). After going Deploy my model to Azure Analysis services and I see 43 rows transferred. PowerBI after refresh see 43 rows too (correct should by 42). When I add column or measure Doplay is success. It is problem only with "refresh" data. Any idea, please?

Hi @JakubBalaj

check processing options of your deployment. If you want to refresh data as well as they changed, you can't use default processing option "default". This doesn't check for changes in data. Use process full instead, or don't process when deploy, rather repocess SSAS db manually afterwards

 

Jiri

 

 

 

@jirineoral

Thank you Jiri, It was helpful 🙂

 

Best regards,

 

Jakub

Suppose i have connect to data source via direct query mode does it effect the underlying data source if some changes are made in Power BI Desktop data in query editor. Please guide....

All new to this.  Played with it before, but now would like to use it at my work.  We use Office 365 so our soruce file (a very basic Excel sheet) is saved in the Cloud. 

What I have done to begin with is:

Use the version of the file on my desktop

Created a report on Power BI Desktop

Saved file

Exported to web based Power BI

I can now see it on the app on my phone (great)

 

It is a database of sales per customer.

This grows on a daily basis.

The report a a basic filter to show sales for a specific customer on my phone.

How do I make this "live", so when a marketer visits a customer and he goes to his app, he has the latest data?

 

As far as I understand from playing around, I will have to go to the Desktop version, updata data, then publish again. (on a daily basis?)

 

Any suggestions would be appreciated.

 

Regards

Hi @Chris_Botha

Excel doesn't support live connection. So you can use just import. If the document is stored in onedrive, you won't need gateway for automated refresh. So you can just set up refreshes (up to 8 per day during working hours). This won't make it real time but at least near real time

 

Jiri

Thank you so much for answering my question.

 

Two more: 

If my file is in Sharepoint on Office 365, can I link it to Power BI - what I mean is, if I can link it to Power BI on the web (not doing it on desktop first and then publish), can I not then have some sort of updating automation?

 

Secondly, you mention a refresh setting, where do I find this?

 

Regards

 

 Any reply to above questions?

Appreciated

@Chris_Botha

refresh options are in dataset settings in your workspace. If you click on workspace, you should get to page where you can see multiple tabs. Dashboards, Reports, Files, Datasets.

Check datasets option.

Jiri