cache tables in apache spark sql

Question

I was going through the documentation of Apache Spark. I couldn't understand this "Chaching tables using the in-memory columnar format"? Help needed.
Thanks in advance.

Answer 1

Caching the tables puts the whole table in memory as spark works on the principle of lazy evaluation. So all the transformations to be done are done when the data is transferred to memory, only the final action requires the data to be retrieved and follow the smart path ie. DAG. 

There's a setting for this too

spark.sql.inMemoryColumnarStorage.compressed = true