Forum Discussion
Paginated report with multi value filter data load service error
I created a direct query model in powerbi that connects with Amazon redshift view to pull data.on top of this model I created a paginated report and created a multivalue filter that pulls data based on user input.
My paginated report is accepting the inputs and pulls data as expected only when the data row count is less than 1000000 without applying any multivalue filter. I am under impression it's enough if the count after applying the user input if it returns 1000000 it's enough . We have billions of rows but after applying the user input it becomes few rows. How does the paginated query works does it not filter values from db itself using user applied inputs ?
Hi Lancy_vini
The available documentation for Paginated Reports provides guidance on the overall workflow, including report authoring, deployment, data connectivity and rendering scenarios.
To help visualize the process, please refer to the official documentation below :
Create and Use the Paginated Report Visual - Power BI | Microsoft Learn
What Are Paginated Reports in Power Bi? - Power BI | Microsoft Learn
9 Replies
Hello Lancy_vini
Please check the below point
Issue:- Power BI does not always filter data first in Redshift
- System scans large data (millions/billions rows)
- Data is processed before filtering
- Causes data load service error
Solution
- Push filter to database (use SQL / stored procedure)
- Avoid large multi-value IN filters
- Reduce data at source (use views / pre-filters)
Thank you
I am trying to understand the architectural flow of powerbi paginated report with direct query model and multivalue parameter for my better understanding
Hi Lancy_vini
The architectural flow in this scenario is:
Paginated Report → Semantic Model (DirectQuery) → DAX Query Engine → DirectQuery Translation → Amazon Redshift.
It highlights how multi-value parameters are passed into the semantic model, converted into DAX, and then translated into SQL for Redshift. The caution note at the bottom reminds that large intermediate datasets may still be processed before filters are fully applied, depending on query folding and model complexity.
