Forum Discussion
Fabric Data Agent Only Returns 25 Rows
- 9 months ago
Hi margaretaak , Thank you for reaching out to the Microsoft Community Forum.
You’re right, the Fabric Data Agent isn’t meant to act like a full data endpoint. It’s designed for conversational insights, not for returning complete datasets. As a result, it automatically limits or summarizes rows in its chat responses to keep things concise and performant.
If you need to visualize all 36 rows, the best approach is to have the agent store your query results in a Lakehouse table or dataset, then connect Quickchart or Copilot Studio directly to that source so you can work with the full dataset.
Have you added any custom instructions or example queries? If so check those in the first instance.
- margaretaak10 months agoFrequent Visitor
Dear KevinChant
Yes, I have added this instruction "do not limit the number of rows returned, return all rows". For example queries, I have added the query I used in SQL Analytics endpoint but it does not change anything. What should I do?
Regards,
Margaretaak
- v-hashadapu10 months agoCommunity Support
Hi margaretaak , Thank you for reaching out to the Microsoft Community Forum.
The difference you’re seeing isn’t caused by your SQL query, it’s because of how Fabric Data Agent handles query responses. The SQL Analytics endpoint will always return all rows you query (in your case, 36 rows) but the Data Agent isn’t designed to return full data tables in its conversational responses. It automatically applies a row limit, typically around 25 to keep interactions concise and performant. Adding natural-language instructions like do not limit the number of rows returned doesn’t override that internal cap.
The only way to get all 36 rows visible for visualization in Copilot Studio with Quickchart is to avoid having the Data Agent directly return the data. Instead, tell it to store the query result in a Lakehouse table or dataset and then have Quickchart connect to that dataset or to the SQL Analytics endpoint directly. This approach ensures the visualization uses the complete result set and not the truncated conversational response.
Query the SQL analytics endpoint of your SQL database - Microsoft Fabric | Microsoft Learn
Fabric data agent creation (preview) - Learn how to create a Fabric data agent | Microsoft Learn