SAP BW Performance

[Syndicate_Admin]

Hello everyone

I am making an SAP Business Warehouse Application Server connection and setting incremental update for some of the tables used. I'm running into performance issues.

First, I saw that in the resulting native query a crossjoin concatenation appears between all the selected dimensions.

I have reviewed in the initial configuration and the default parameters have been left after not having obtained a performance improvement (BasXmlGzip and batch size 50000).

Once uploaded to Power BI Service, an incremental refresh policy of the last two months is being established, obtaining an update time of 8 minutes in the case of forcing the update and about 16 in case of performing a scheduled update. Considering the size of the table, the time should be much less, how can I improve the performance? Is the structure of the native query normal?

Thank you.

[Anonymous · Accepted solution]

Hi @Anonymous ,

When you configure Power BI for incremental refresh, the first load will process all the data to create the initial partitions. This process might take longer than a regular refresh because it needs to create and store the partitions for future incremental refreshes. Once the initial load is complete, subsequent refreshes should be faster as they will only process the new or changed data. If that is not first time, and the refresh time is longer than a schedule refresh. It may caused by various factors such as query complexity, lack of query folding, or issues with the data source itself. To improve the performance of incremental refresh, you can try the following:Optimize your Power

1. Query M (Mashup) queries by simplifying them and pushing complexity to the data source side if possible.
2. Ensure that query folding is happening by using Mashup functions that support query folding as early as possible in the steps.

        Incremental refresh&query folding | Microsoft Power BI Blog | Microsoft Power BI

      3. Check if the data source's tables have proper indexing and statistics.

      4. If the data source allows, use profiling tools to analyze the queries received on the data source side.

          Using the data profiling tools - Power Query | Microsoft Learn

And you can refer the following official documentation to make troubleshooting and try to get a solution.

Troubleshoot incremental refresh and real-time data in Power BI - Power BI | Microsoft Learn

Additionally, you can consider scheduling the dataset refresh at a different time when there's less overhead or try using premium capacity for more resources.

What is Power BI Premium? - Power BI | Microsoft Learn

Best Regards