What are the trade-offs between different storage formats in a data lake environment, and how do you choose for Cloudera analytics use cases?
