Warehouses, SQL editor, parameters, and Assistant

The warehouse executes the SQL. Dashboards, alerts, and Genie are only as healthy as the warehouse they bind to (running, sized, permitted).

The exam states SQL follows ANSI. Prefer portable joins, GROUP BY, windows, CASE. Databricks extras (time travel, UC names, IDENTIFIER) appear when the question is clearly platform-specific.

Named parameters (:start_date) in saved queries and dashboard datasets push filters into SQL so the warehouse scans less. A date widget’s job is typically to supply values used in a WHERE clause, not merely to label the x-axis.

sqlSELECT order_date, SUM(amount) AS revenue
FROM catalog.gold.fact_orders
WHERE order_date BETWEEN :start_date AND :end_date
GROUP BY order_date;

Use `/explain` when a complex query returns the wrong result and you need a walkthrough. Use generate to draft, then verify.

SQL editor · SQL warehouse