Pyspark pivot on multiple columns without join or reduce

Question

I need to count number of appearances for each value in multiple columns and group by the output by the same fields for all of them. For example:

+-------------+-