even though that isn't the pre-selected default anymore.

Databricks runtime 6.4 ist the last one to feature Spark 2.x
Spark 3.x comes with some changes that would require adapting
the exercise notebook:

- pattern language for date formats changed

- `spark.conf.get("spark.sql.files.maxPartitionBytes")` now returns
  '134217728b' instead of just
  '134217728' (The 'b' is Spark's unit indicator for "bytes")