Deriving Number of Partitions in Spark SQL

Question

We deal with large datasets in our project and hence we encounter performance issues in our Spark SQL queries (especially in Joins) quite often. Hence, I am trying to figure out