When would you use cache versus persist in Spark?

In Spark, cache() and persist() are used to store RDDs or DataFrames in memory for reuse across actions. cache() is a shorthand for persist(StorageLevel.MEMORY_ONLY) and is convenient when you only need in‑memory storage and are willing to recompute if memory is insufficient. persist() gives you fine‑grained control over the storage level, allowing you to choose MEMORY_AND_DISK, DISK_ONLY, or even OFF_HEAP. It also lets you change the level after the first action. Use cache() for quick, in‑memory caching of small to medium datasets. Use persist() when you need to control spill behavior, keep data on disk, or change the storage level later. In practice, cache() is the default choice for simple caching, while persist() is used for more complex persistence requirements or when you need to avoid recomputation under memory pressure.