casewhen.py

#   Read the source tables in Parquet format
sales_table = spark.read.parquet("./data/sales_parquet")

'''
SELECT seller_id,
       CASE WHEN num_pieces_sold < 30 THEN 'Lower than 30',
            WHEN num_pieces_sold < 60 THEN 'Between 31 and 60'
            WHEN num_pieces_sold < 90 THEN 'Between 61 and 90'
            ELSE 'More than 91' AS sales_bucket
FROM sales_table
'''
sales_table_execution_plan = sales_table.select(
    col("seller_id"),
    when(col("num_pieces_sold") < 30, "Lower than 30").
    when(col("num_pieces_sold") < 60, "Between 31 and 60").
    when(col("num_pieces_sold") < 90, "Between 61 and 90").
    otherwise("More than 91").alias("sales_bucket")
)

sales_table_execution_plan.show()