Spike: consider how to solve caching of the measured DataFrame

[benedeki]

Background

DataFrame which is measured might contain expensive operations.

Feature

Consider how to solve this. Maybe it' would be as easy as adding a call to .cache but maybe something more elaborate.
Also if it would be (just) .cache consider making it possible to switch off - globally, or for that one particular call.

[jakipatryk]

My initial thinking is that it should be the app using Atum doing the cache/persist before saving checkpoint (?) in Atum. But if so it has to be stated explicitly somewhere, maybe even with some warning message if input DF in not persisted