Comparing randsample and binning uniform random

I can produce randomly polled (with replacement) values from vector x with probabilities in vector p using randsample with replacement.

X = randsample(N,K,true,W)

I can equivalently do this by using rand and histcounts (specifying edges) that correspond to the values from a cumulative sum on vector p (assuming p is normalized).

edges = [0 cumsum(p)]; edges(end)=1; % for possible rounding error
Y = rand(1,nsamples);
[~,~,bin] = histcounts(Y, edges);
X = x(bin);

My question is: if one of these is better, which one is?

[Jan]

randsample is nicer, without doubt.

[yeungor]

Thanks, it's more readable, that's for sure.