Performance for a report with big source data

[tomas12344]

HI All

Have a very special report, that uses a query in SAP BW, which is quite big. We already managed to somehow reduce the number of lines, but I think it is still big.

Current status is, that we are able to load 2 years of the data from SAP BW to PBI (impot mode), but there are still lot of rows - 70M rows (the load is quite fast only 2hours). 
The question is, if you can estimate any performance for such big dataset, eventualy what to check? If you have any experience with such big dataset? 

Currently the customer want only some 3 small tables, but in the future they want to extend the report. Report should run in Service, so any suggestion for consideration would be helpful.

[Greg_Deckler · Accepted solution]

@tomas12344 So import mode is going to be the fastest in terms of performance. The actual performance of the report will be mainly controlled by how many visuals you have on a page and how complex your DAX measures are. It is extremely difficult to know any of that until you actually create the report. 

[Anonymous · Accepted solution]

Hi @tomas12344 ,

Microsoft provides a comprehensive guide on optimizing Power BI solutions, including data model optimization, visualization optimization, and environment optimization. Please refer: Optimization guide for Power BI - Power BI | Microsoft Learn

In addition to this, incremental refresh allows you to load and refresh data in increments, significantly reducing load times and resource usage.

If you're not already, consider using Power BI Premium. It provides dedicated capacity which can handle larger datasets and more complex reports than the shared capacity.

Best Regards,

Neeko Tang

If this post  helps, then please consider Accept it as the solution  to help the other members find it more quickly. 

[Anonymous · Accepted solution]

Hi @tomas12344 ,

Microsoft provides a comprehensive guide on optimizing Power BI solutions, including data model optimization, visualization optimization, and environment optimization. Please refer: Optimization guide for Power BI - Power BI | Microsoft Learn

In addition to this, incremental refresh allows you to load and refresh data in increments, significantly reducing load times and resource usage.

If you're not already, consider using Power BI Premium. It provides dedicated capacity which can handle larger datasets and more complex reports than the shared capacity.

Best Regards,

Neeko Tang

If this post  helps, then please consider Accept it as the solution  to help the other members find it more quickly. 

[anupamj · Accepted solution]

Hello @tomas12344, for such an extraction you may have to split the load into multiple packets.

you will find these useful:

1. Guide for high volume data replication from SAP : https://www.youtube.com/playlist?list=PLTum8dvrbVA2nqRs4pUN29XzJ8Qd-OUym
2. Once full extraction is done, scheduled delta capture from SAP : https://www.youtube.com/playlist?list=PLTum8dvrbVA1lguJcC3aX1w4NswhHDn_0

Microsoft recommended architecture for PowerBI / Fabric integration with SAP (See the cached method): 

https://techcommunity.microsoft.com/t5/azure-data-factory-blog/integrate-microsoft-fabric-with-sap-d...

Questions? please reach out.

Anupam

[Greg_Deckler · Accepted solution]

@tomas12344 So import mode is going to be the fastest in terms of performance. The actual performance of the report will be mainly controlled by how many visuals you have on a page and how complex your DAX measures are. It is extremely difficult to know any of that until you actually create the report.