If you want the resulting array to contain exactly n(1) copies of x(1), n(2) copies of x(2), etc. use the repelem function. [Depending on how you want to use the result, you may also want to "shuffle" it using randperm.]
x = 0:5;
freq = [2 3 4 5 6 7];
A = repelem(x, freq);
If instead of frequencies you have probabilities for what each element of the resulting array should be, use discretize.
p = freq./sum(freq);
cumulativeprob = [0 cumsum(p)];
B = discretize(rand(1, 1000), cumulativeprob, x);
Let's check that the proportions of elements in B come close to the probabilities in p.
c = histcounts(B, 'BinEdges', [x x(end)], 'Normalization', 'probability');
[c; p]
The values in c look reasonably close to the probabilities in p for just 1000 samples.
