Waiting 1m is still better than waiting 3m. Disregarding the less significant factors in O(3n) is not always a good idea in practice.
I agree all of this is a pretty useless exercise when you have 3 consecutive maps over 10 or even 1000 items, but this optimization is the reason many FP libraries offer super easy to use pipe/compose functions.
I'd say, if the optimization is that easy, do it — even if the gains small. Without pipe/compose available, I'd be in favor of readability over performance though.
What are you talking about. The time complexity doesn't change by moving everything into the reducer, it's the space complexity that changes because .map creates a new array. It's not 1ms vs 3ms...
You are right that it doesn't take three times as long, my mistake.
But it's never just "1+1+1=3" either.
Especially if the arrays are large, and the operations are simple, then the overhead of copying the array doesn't just affect memory use, it also takes longer.
We're a place where coders share, stay up-to-date and grow their careers.
We strive for transparency and don't collect excess data.