Most PySpark Tuning Advice Starts in the Wrong Place

PySpark runs DataFrame code in the JVM at native speed, and almost everything that makes it slow happens where data crosses into Python. Ordered...