df-window-func-rank-topN.py

spark.sql("""
WITH CALCULATED (
    SELECT 
        CAST(event_time AS DATE) as event_date,
        brand, 
        sum(price) as price_sum
        
    FROM PURCHASE
    
    WHERE 
        brand IS NOT NULL 
        
    GROUP BY 1, 2
), 

RANKED (
    SELECT 
        event_date,
        brand, 
        price_sum, 
        rank() OVER (PARTITION BY event_date ORDER BY price_sum DESC) as rank
        
    FROM CALCULATED
)

SELECT *
FROM RANKED
WHERE rank <= 3
ORDER BY event_date ASC, rank ASC
""").show(truncate=False)