How would you use partitioning to improve query performance on large historical datasets in TaskRabbit’s warehouse?
