Hi Jacek,
I haven't played with 2.1.0 yet, so not sure how much more optimized Window
functions are compared to 1.6 and 2.0.
However, one thing I do see in the self-join is a broadcast. So there's going
to be a need broadcast the results of the groupBy out to the executors before
it can do
Hi,
While playing around with Spark 2.1.0-SNAPSHOT (built today) and
explain'ing two queries with WindowSpec and inner join I found the
following plans and am wondering if you could help me to judge which
query could be faster.
What else would you ask for to be able to answer the question of one
