How do you design a partitioning and clustering strategy for massive tables stored in cloud object storage?
