I am currently using FPGrowth from Spark + MLLib, version 2.12.
It currently works as follows:
// 1. Create datase