databricks-aibi-dashboards

The "no selected fields to visualize" error in Databricks AI/BI dashboards occurs when dataset query field names do not exactly match widget encodings fieldName properties. Enforcing strict dataset-to-widget contracts ensures exact name matching between Spark SQL query outputs and widget configurations to resolve this rendering failure.

Safely deploying Databricks AI/BI dashboards requires authoring validated Spark SQL datasets, executing mandatory query testing via execute_sql, assembling dashboard JSON with strict widget contracts, and using manage_dashboard with action create_or_update for controlled releases.

Yes, mandatory execute_sql testing for every Spark SQL query is required before adding it to an AI/BI dashboard. This validation workflow prevents broken visualizations by verifying schema output and ensuring exact fieldName matching between datasets and widget encodings.

Databricks AI/BI dashboard widgets must adhere to GRID_V1 layout rules, requiring placement within a 12-column grid structure without any gaps. Correct widget versioning and proper filter widget types are also necessary for reliable dashboard rendering.

You can update existing Databricks AI/BI dashboards without recreating them by using the manage_dashboard deployment function with the create_or_update action. This allows seamless updates to KPI counters, tables, charts, and global or page-level filters while maintaining widget validation.

Databricks AI/BI dashboards break when adding filters if the filter widget types are incorrect or if there are mismatched field contracts between the filter and the dataset. Applying correct widget versions and exact fieldName matching ensures filters function properly across the dashboard.