Is usage of
join()as finally shown by @iwvelandoexperimental.join()as shown by @rickspencer3 r
really the proper solution for arithmetic with fields of a single measurement ?
My understanding is that pivot() is best practices method, see
With schema.fieldsAsCols() doing field arithmetic becomes really easy and compact, as explained by @Anaisdg .
Beyond that, join() seems to have a performance problem, see
Poor performance for join(): cpu and memory grows quadratically with row count
and experimental.join() seems to have a stability problem, see
Tried experimental.join(): good performance, but also "panic: unknown type invalid"
