So I hear a lot about how TopN and TimeSeries are way faster than GroupBy.
However, there is no technical details on ‘what’ exactly makes that happen.
The most I have heard is that TopN’s use’ approximation’ (that word by itself is not enough detail).
GroupBys are much easier to program against and a lot of times do the job of multiple TopN/TimeSeries statements.
Without any details about what makes GroupBys slower, it is almost impossible to currently weigh your choices of easier programming vs faster UI.
Can someone please provide some technical details on the internals of GroupBys?
A simple useful example might be to show how a GroupBy might go about processing data vs a TopN.