Add Pyspark.pandas to benchmark

[TjommeVergauwen]

Are there any plans to add Pyspark.pandas to the benchmark?

[jangorecki]

Do you expect to have different performance than pyspark.sql? Do you think it will be faster or slower?
I think it make sense to keep only one of them rather than maintaining both. Running each solution costs couple hours of high spec machine, so I would avoid benchmarking spark interfaces (SQL/pandas) and focus on the engine. I am sure they share the same spark engine.