The two notebooks implement the same application in two alternative ways:
- Exam20260901SparkRDD.ipynb: RDD-based solution. Part 2 uses the static-window
  pattern based on flatMap + reduceByKey shown in the course examples;
- Exam20260901SparkSQL.ipynb: Spark SQL-based solution.

The application receives all three input paths. Customers.txt is deliberately
not read because neither requested result uses customer profile attributes.

The folders data/ and datacsv/ contain the same coherent example dataset,
without and with headers, respectively. The expected results are reported in
expected_output1.txt and expected_output2.txt. Spark does not guarantee the
order of output rows.

For Part 2, the example dataset follows the formal task specification: a month
qualifies for a customer when purchases in that month involve at least 10
distinct categories. Two consecutive qualifying months produce one line with
the customer identifier and the numeric starting month.
