Maintain Analytics Data

For: Analytics Writers and Publishers

Use Writer access for row-level data changes and Publisher access for structural changes. Both require an MCP token with lab.analytics.write.

Common tasks

• Add new rows to an existing table.
• Update or replace existing data.
• Load another CSV batch.
• Review load status and imported row counts.
• Modify a supported table structure.
• Resolve failed or partially completed loads.
• Verify data after maintenance.

Load a CSV file

1. Confirm the Analytics Project, dataset, table, and expected columns.
2. Decide whether the load should append rows or replace the existing table data.
3. Attach the CSV file and ask the AI agent to inspect its headers and compatibility.
4. Review the target and load mode before continuing.
5. After the load, check the reported row count, columns, status, and audit ID.
6. Run a read-only verification query against the updated table.

Example:

Load the attached CSV into study_data.measurements in the Respiratory Cohort Analytics Project. Use append mode. Confirm the target table and columns before loading, then report the load status, row count, and audit ID.

Change a table structure

Ask the AI agent to show the existing structure and proposed change before adding, renaming, or removing a column. A rename or data-type change can affect saved queries and views, so identify those dependencies first.

Caution

Removing a column or table is destructive. Confirm the exact Analytics Project and dataset object before approving the change.