File upload (CSV / Excel)
Upload a CSV or Excel file directly as a dataset. This is the simplest way to analyze a one-off file with no external storage or credentials.
When to use
- A single spreadsheet or CSV export you want to analyze immediately.
- Ad-hoc data that does not live in a database or object storage.
For recurring or large data, a database or object-storage connector is more appropriate; an upload is a point-in-time copy.
How it works
Choose upload when adding a dataset and select the file. Datyo parses it, detects columns and their types automatically, and reads a sample so the Agent can reason about it. Columns and types are inferred from the content; a clear header row improves detection.
Supported formats
- CSV (comma-separated values).
- Excel workbooks (
.xlsx).
Capabilities
- Execution model: Federated. Uploaded data can be joined with datasets from other connectors.
- Access: Read-only analysis of the uploaded copy.
- Point-in-time: An upload is a snapshot; to analyze newer data, upload again or use a live connector.
Troubleshooting
- Columns or types are wrong — ensure the first row is a header and that each column has consistent value types; re-upload after correcting the file.
- Encoding or delimiter issues — save the CSV as UTF-8; irregular delimiters can cause misparsing.
Improve answer accuracy
- Add column descriptions so the Agent maps business terms to the right columns.