Connect Excel to Delta Forge through the ODBC driver via the Get Data > From ODBC route.
Data ribbon > Get Data > From Other Sources > From ODBC## Excel on Windows 1. **Data** ribbon -> **Get Data** -> **From Other Sources** -> **From ODBC**. 2. Pick `DeltaForge` from the dropdown. 3. Excel uses the same ODBC entry points as Power BI (both consume the Power Query engine under the hood). Anything that works in Power BI Desktop works here. 4. The navigator browses zones / schemas / tables. Pick one, **Load** to a worksheet or **Transform Data** to open the Power Query editor first. ## Excel on macOS ODBC support on Excel for Mac goes through **Microsoft Query**, which is materially less capable than the Windows path: - The connection dialog only accepts iODBC DSNs. - The navigator does not browse the catalog; you must type the SQL directly. - Pivot table refresh works but is slow. For non-trivial work on macOS, prefer one of: - Pull data with Python or R, write to a `.csv` or `.xlsx`, open in Excel. - Use the Delta Forge GUI to export to a file and open from there. - Move the analysis to a Windows VM where the full Power Query path is available. ## Refresh Excel saves the connection in the workbook. Closing and reopening rehydrates the connection from the DSN; refresh runs every query again. Power Query queries can be promoted to **Connection-only** to keep them in the workbook without materialising the result, which is useful for staging queries that feed a final model. ## Performance Excel materialises the entire result set into worksheet cells. For large queries, push as much filtering as possible to the SQL using **Custom SQL**: ```sql SELECT customer_id, sum(amount) AS total_amount FROM analytics.marts.fact_orders WHERE order_date >= DATE '2026-01-01' GROUP BY customer_id ``` is dramatically faster than loading `fact_orders` whole and pivoting in Excel.
# DSN-based connect
DSN=DeltaForge;Uid=alice@example.com# Driver-string form
Driver={DeltaForge ODBC Driver};Server=https://df.example.com;Uid=alice@example.com;Pwd=hunter2