You possess historical data spanning the last three years stored in BigQuery, alongside a data pipeline that updates this BigQuery dataset with new daily data. Observations have shown that when the Data Science team executes a query filtered by a date column and constrained to a 30-90 day range, the entire table is unintentionally scanned. This extensive table scanning has led to an unexpected surge in costs. To address this, your goal is to find a cost-effective solution that still allows for efficient SQL querying. What steps should you take?