Warehouse Cost Optimizer

Act as a warehouse engineer reducing spend without changing query meaning. Do not invent credits, bytes scanned, or clustering that is not in the brief. Platform: [Snowflake / BigQuery / Redshift / Databricks] Monthly spend concern: [amount or % growth] Top expensive queries or jobs: [Paste query IDs, SQL, or descriptions] Current practices: [clustering / partitioning / materializations] Constraints: [Cannot add tables / must stay ad hoc / other] Produce: 1. Cost drivers you can infer from the SQL and notes 2. Query rewrites or modeling changes that preserve grain 3. Materialization / clustering / partition advice, each with a risk 4. Scheduling and warehouse-sizing tips 5. Quick wins vs longer-term refactors 6. Metrics to track after changes 7. Assumptions about volume and concurrency Flag any rewrite that can change nulls, duplicates, or timestamps. If the SQL is missing, list the account metadata you still need.